1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
3 @c Free Software Foundation, Inc.
4 @c This is part of the GCC manual.
5 @c For copying conditions, see the file gcc.texi.
12 @c man begin COPYRIGHT
13 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
14 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
15 Free Software Foundation, Inc.
17 Permission is granted to copy, distribute and/or modify this document
18 under the terms of the GNU Free Documentation License, Version 1.3 or
19 any later version published by the Free Software Foundation; with the
20 Invariant Sections being ``GNU General Public License'' and ``Funding
21 Free Software'', the Front-Cover texts being (a) (see below), and with
22 the Back-Cover Texts being (b) (see below). A copy of the license is
23 included in the gfdl(7) man page.
25 (a) The FSF's Front-Cover Text is:
29 (b) The FSF's Back-Cover Text is:
31 You have freedom to copy and modify this GNU Manual, like GNU
32 software. Copies published by the Free Software Foundation raise
33 funds for GNU development.
35 @c Set file name and title for the man page.
37 @settitle GNU project C and C++ compiler
39 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
40 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
41 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
42 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
43 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
44 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
45 [@option{-o} @var{outfile}] [@@@var{file}] @var{infile}@dots{}
47 Only the most useful options are listed here; see below for the
48 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
51 gpl(7), gfdl(7), fsf-funding(7),
52 cpp(1), gcov(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
53 and the Info entries for @file{gcc}, @file{cpp}, @file{as},
54 @file{ld}, @file{binutils} and @file{gdb}.
57 For instructions on reporting bugs, see
61 See the Info entry for @command{gcc}, or
62 @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}},
63 for contributors to GCC@.
68 @chapter GCC Command Options
69 @cindex GCC command options
70 @cindex command options
71 @cindex options, GCC command
73 @c man begin DESCRIPTION
74 When you invoke GCC, it normally does preprocessing, compilation,
75 assembly and linking. The ``overall options'' allow you to stop this
76 process at an intermediate stage. For example, the @option{-c} option
77 says not to run the linker. Then the output consists of object files
78 output by the assembler.
80 Other options are passed on to one stage of processing. Some options
81 control the preprocessor and others the compiler itself. Yet other
82 options control the assembler and linker; most of these are not
83 documented here, since you rarely need to use any of them.
85 @cindex C compilation options
86 Most of the command line options that you can use with GCC are useful
87 for C programs; when an option is only useful with another language
88 (usually C++), the explanation says so explicitly. If the description
89 for a particular option does not mention a source language, you can use
90 that option with all supported languages.
92 @cindex C++ compilation options
93 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
94 options for compiling C++ programs.
96 @cindex grouping options
97 @cindex options, grouping
98 The @command{gcc} program accepts options and file names as operands. Many
99 options have multi-letter names; therefore multiple single-letter options
100 may @emph{not} be grouped: @option{-dv} is very different from @w{@samp{-d
103 @cindex order of options
104 @cindex options, order
105 You can mix options and other arguments. For the most part, the order
106 you use doesn't matter. Order does matter when you use several
107 options of the same kind; for example, if you specify @option{-L} more
108 than once, the directories are searched in the order specified. Also,
109 the placement of the @option{-l} option is significant.
111 Many options have long names starting with @samp{-f} or with
112 @samp{-W}---for example,
113 @option{-fmove-loop-invariants}, @option{-Wformat} and so on. Most of
114 these have both positive and negative forms; the negative form of
115 @option{-ffoo} would be @option{-fno-foo}. This manual documents
116 only one of these two forms, whichever one is not the default.
120 @xref{Option Index}, for an index to GCC's options.
123 * Option Summary:: Brief list of all options, without explanations.
124 * Overall Options:: Controlling the kind of output:
125 an executable, object files, assembler files,
126 or preprocessed source.
127 * Invoking G++:: Compiling C++ programs.
128 * C Dialect Options:: Controlling the variant of C language compiled.
129 * C++ Dialect Options:: Variations on C++.
130 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
132 * Language Independent Options:: Controlling how diagnostics should be
134 * Warning Options:: How picky should the compiler be?
135 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
136 * Optimize Options:: How much optimization?
137 * Preprocessor Options:: Controlling header files and macro definitions.
138 Also, getting dependency information for Make.
139 * Assembler Options:: Passing options to the assembler.
140 * Link Options:: Specifying libraries and so on.
141 * Directory Options:: Where to find header files and libraries.
142 Where to find the compiler executable files.
143 * Spec Files:: How to pass switches to sub-processes.
144 * Target Options:: Running a cross-compiler, or an old version of GCC.
145 * Submodel Options:: Specifying minor hardware or convention variations,
146 such as 68010 vs 68020.
147 * Code Gen Options:: Specifying conventions for function calls, data layout
149 * Environment Variables:: Env vars that affect GCC.
150 * Precompiled Headers:: Compiling a header once, and using it many times.
156 @section Option Summary
158 Here is a summary of all the options, grouped by type. Explanations are
159 in the following sections.
162 @item Overall Options
163 @xref{Overall Options,,Options Controlling the Kind of Output}.
164 @gccoptlist{-c -S -E -o @var{file} -combine -no-canonical-prefixes @gol
165 -pipe -pass-exit-codes @gol
166 -x @var{language} -v -### --help@r{[}=@var{class}@r{[},@dots{}@r{]]} --target-help @gol
167 --version -wrapper@@@var{file} -fplugin=@var{file} -fplugin-arg-@var{name}=@var{arg} @gol
168 -fdump-ada-spec@r{[}-slim@r{]}}
170 @item C Language Options
171 @xref{C Dialect Options,,Options Controlling C Dialect}.
172 @gccoptlist{-ansi -std=@var{standard} -fgnu89-inline @gol
173 -aux-info @var{filename} @gol
174 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
175 -fhosted -ffreestanding -fopenmp -fms-extensions @gol
176 -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol
177 -fallow-single-precision -fcond-mismatch -flax-vector-conversions @gol
178 -fsigned-bitfields -fsigned-char @gol
179 -funsigned-bitfields -funsigned-char}
181 @item C++ Language Options
182 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
183 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
184 -fconserve-space -ffriend-injection @gol
185 -fno-elide-constructors @gol
186 -fno-enforce-eh-specs @gol
187 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
188 -fno-implicit-templates @gol
189 -fno-implicit-inline-templates @gol
190 -fno-implement-inlines -fms-extensions @gol
191 -fno-nonansi-builtins -fnothrow-opt -fno-operator-names @gol
192 -fno-optional-diags -fpermissive @gol
193 -fno-pretty-templates @gol
194 -frepo -fno-rtti -fstats -ftemplate-depth=@var{n} @gol
195 -fno-threadsafe-statics -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
196 -fno-default-inline -fvisibility-inlines-hidden @gol
197 -fvisibility-ms-compat @gol
198 -Wabi -Wconversion-null -Wctor-dtor-privacy @gol
199 -Wnoexcept -Wnon-virtual-dtor -Wreorder @gol
200 -Weffc++ -Wstrict-null-sentinel @gol
201 -Wno-non-template-friend -Wold-style-cast @gol
202 -Woverloaded-virtual -Wno-pmf-conversions @gol
205 @item Objective-C and Objective-C++ Language Options
206 @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling
207 Objective-C and Objective-C++ Dialects}.
208 @gccoptlist{-fconstant-string-class=@var{class-name} @gol
209 -fgnu-runtime -fnext-runtime @gol
210 -fno-nil-receivers @gol
211 -fobjc-call-cxx-cdtors @gol
212 -fobjc-direct-dispatch @gol
213 -fobjc-exceptions @gol
215 -freplace-objc-classes @gol
218 -Wassign-intercept @gol
219 -Wno-protocol -Wselector @gol
220 -Wstrict-selector-match @gol
221 -Wundeclared-selector}
223 @item Language Independent Options
224 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
225 @gccoptlist{-fmessage-length=@var{n} @gol
226 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]} @gol
227 -fdiagnostics-show-option}
229 @item Warning Options
230 @xref{Warning Options,,Options to Request or Suppress Warnings}.
231 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
232 -w -Wextra -Wall -Waddress -Waggregate-return -Warray-bounds @gol
233 -Wno-attributes -Wno-builtin-macro-redefined @gol
234 -Wc++-compat -Wc++0x-compat -Wcast-align -Wcast-qual @gol
235 -Wchar-subscripts -Wclobbered -Wcomment @gol
236 -Wconversion -Wcoverage-mismatch -Wcpp -Wno-deprecated @gol
237 -Wno-deprecated-declarations -Wdisabled-optimization @gol
238 -Wno-div-by-zero -Wempty-body -Wenum-compare -Wno-endif-labels @gol
239 -Werror -Werror=* @gol
240 -Wfatal-errors -Wfloat-equal -Wformat -Wformat=2 @gol
241 -Wno-format-contains-nul -Wno-format-extra-args -Wformat-nonliteral @gol
242 -Wformat-security -Wformat-y2k @gol
243 -Wframe-larger-than=@var{len} -Wjump-misses-init -Wignored-qualifiers @gol
244 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
245 -Winit-self -Winline @gol
246 -Wno-int-to-pointer-cast -Wno-invalid-offsetof @gol
247 -Winvalid-pch -Wlarger-than=@var{len} -Wunsafe-loop-optimizations @gol
248 -Wlogical-op -Wlong-long @gol
249 -Wmain -Wmissing-braces -Wmissing-field-initializers @gol
250 -Wmissing-format-attribute -Wmissing-include-dirs @gol
252 -Wno-multichar -Wnonnull -Wno-overflow @gol
253 -Woverlength-strings -Wpacked -Wpacked-bitfield-compat -Wpadded @gol
254 -Wparentheses -Wpedantic-ms-format -Wno-pedantic-ms-format @gol
255 -Wpointer-arith -Wno-pointer-to-int-cast @gol
256 -Wredundant-decls @gol
257 -Wreturn-type -Wsequence-point -Wshadow @gol
258 -Wsign-compare -Wsign-conversion -Wstack-protector @gol
259 -Wstrict-aliasing -Wstrict-aliasing=n @gol
260 -Wstrict-overflow -Wstrict-overflow=@var{n} @gol
261 -Wsuggest-attribute=@r{[}const@r{|}pure@r{]} @gol
262 -Wswitch -Wswitch-default -Wswitch-enum -Wsync-nand @gol
263 -Wsystem-headers -Wtrigraphs -Wtype-limits -Wundef -Wuninitialized @gol
264 -Wunknown-pragmas -Wno-pragmas @gol
265 -Wunsuffixed-float-constants -Wunused -Wunused-function @gol
266 -Wunused-label -Wunused-parameter -Wno-unused-result -Wunused-value -Wunused-variable @gol
267 -Wunused-but-set-parameter -Wunused-but-set-variable -Wvariadic-macros -Wvla @gol
268 -Wvolatile-register-var -Wwrite-strings}
270 @item C and Objective-C-only Warning Options
271 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
272 -Wmissing-parameter-type -Wmissing-prototypes -Wnested-externs @gol
273 -Wold-style-declaration -Wold-style-definition @gol
274 -Wstrict-prototypes -Wtraditional -Wtraditional-conversion @gol
275 -Wdeclaration-after-statement -Wpointer-sign}
277 @item Debugging Options
278 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
279 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
280 -fdbg-cnt-list -fdbg-cnt=@var{counter-value-list} @gol
281 -fdump-noaddr -fdump-unnumbered -fdump-unnumbered-links @gol
282 -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
283 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
284 -fdump-ipa-all -fdump-ipa-cgraph -fdump-ipa-inline @gol
285 -fdump-statistics @gol
287 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
288 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
289 -fdump-tree-cfg -fdump-tree-vcg -fdump-tree-alias @gol
291 -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol
292 -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol
293 -fdump-tree-gimple@r{[}-raw@r{]} -fdump-tree-mudflap@r{[}-@var{n}@r{]} @gol
294 -fdump-tree-dom@r{[}-@var{n}@r{]} @gol
295 -fdump-tree-dse@r{[}-@var{n}@r{]} @gol
296 -fdump-tree-phiprop@r{[}-@var{n}@r{]} @gol
297 -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol
298 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
299 -fdump-tree-copyrename@r{[}-@var{n}@r{]} @gol
300 -fdump-tree-nrv -fdump-tree-vect @gol
301 -fdump-tree-sink @gol
302 -fdump-tree-sra@r{[}-@var{n}@r{]} @gol
303 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
304 -fdump-tree-fre@r{[}-@var{n}@r{]} @gol
305 -fdump-tree-vrp@r{[}-@var{n}@r{]} @gol
306 -ftree-vectorizer-verbose=@var{n} @gol
307 -fdump-tree-storeccp@r{[}-@var{n}@r{]} @gol
308 -fdump-final-insns=@var{file} @gol
309 -fcompare-debug@r{[}=@var{opts}@r{]} -fcompare-debug-second @gol
310 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
311 -feliminate-unused-debug-symbols -femit-class-debug-always @gol
312 -fenable-icf-debug @gol
313 -fmem-report -fpre-ipa-mem-report -fpost-ipa-mem-report -fprofile-arcs @gol
314 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
315 -fsel-sched-verbose -fsel-sched-dump-cfg -fsel-sched-pipelining-verbose @gol
316 -ftest-coverage -ftime-report -fvar-tracking @gol
317 -fvar-tracking-assignments -fvar-tracking-assignments-toggle @gol
318 -g -g@var{level} -gtoggle -gcoff -gdwarf-@var{version} @gol
319 -ggdb -gstabs -gstabs+ -gstrict-dwarf -gno-strict-dwarf @gol
320 -gvms -gxcoff -gxcoff+ @gol
321 -fno-merge-debug-strings -fno-dwarf2-cfi-asm @gol
322 -fdebug-prefix-map=@var{old}=@var{new} @gol
323 -femit-struct-debug-baseonly -femit-struct-debug-reduced @gol
324 -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]} @gol
325 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
326 -print-multi-directory -print-multi-lib -print-multi-os-directory @gol
327 -print-prog-name=@var{program} -print-search-dirs -Q @gol
328 -print-sysroot -print-sysroot-headers-suffix @gol
329 -save-temps -save-temps=cwd -save-temps=obj -time@r{[}=@var{file}@r{]}}
331 @item Optimization Options
332 @xref{Optimize Options,,Options that Control Optimization}.
334 -falign-functions[=@var{n}] -falign-jumps[=@var{n}] @gol
335 -falign-labels[=@var{n}] -falign-loops[=@var{n}] -fassociative-math @gol
336 -fauto-inc-dec -fbranch-probabilities -fbranch-target-load-optimize @gol
337 -fbranch-target-load-optimize2 -fbtr-bb-exclusive -fcaller-saves @gol
338 -fcheck-data-deps -fconserve-stack -fcprop-registers -fcrossjumping @gol
339 -fcse-follow-jumps -fcse-skip-blocks -fcx-fortran-rules -fcx-limited-range @gol
340 -fdata-sections -fdce -fdce @gol
341 -fdelayed-branch -fdelete-null-pointer-checks -fdse -fdse @gol
342 -fearly-inlining -fipa-sra -fexpensive-optimizations -ffast-math @gol
343 -ffinite-math-only -ffloat-store -fexcess-precision=@var{style} @gol
344 -fforward-propagate -ffunction-sections @gol
345 -fgcse -fgcse-after-reload -fgcse-las -fgcse-lm @gol
346 -fgcse-sm -fif-conversion -fif-conversion2 -findirect-inlining @gol
347 -finline-functions -finline-functions-called-once -finline-limit=@var{n} @gol
348 -finline-small-functions -fipa-cp -fipa-cp-clone -fipa-matrix-reorg -fipa-pta @gol
349 -fipa-profile -fipa-pure-const -fipa-reference -fipa-struct-reorg @gol
350 -fira-algorithm=@var{algorithm} @gol
351 -fira-region=@var{region} -fira-coalesce @gol
352 -fira-loop-pressure -fno-ira-share-save-slots @gol
353 -fno-ira-share-spill-slots -fira-verbose=@var{n} @gol
354 -fivopts -fkeep-inline-functions -fkeep-static-consts @gol
355 -floop-block -floop-interchange -floop-strip-mine -fgraphite-identity @gol
356 -floop-parallelize-all -flto -flto-compression-level -flto-report -fltrans @gol
357 -fltrans-output-list -fmerge-all-constants -fmerge-constants -fmodulo-sched @gol
358 -fmodulo-sched-allow-regmoves -fmove-loop-invariants -fmudflap @gol
359 -fmudflapir -fmudflapth -fno-branch-count-reg -fno-default-inline @gol
360 -fno-defer-pop -fno-function-cse -fno-guess-branch-probability @gol
361 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
362 -fno-sched-interblock -fno-sched-spec -fno-signed-zeros @gol
363 -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
364 -fomit-frame-pointer -foptimize-register-move -foptimize-sibling-calls @gol
365 -fpeel-loops -fpredictive-commoning -fprefetch-loop-arrays @gol
366 -fprofile-correction -fprofile-dir=@var{path} -fprofile-generate @gol
367 -fprofile-generate=@var{path} @gol
368 -fprofile-use -fprofile-use=@var{path} -fprofile-values @gol
369 -freciprocal-math -fregmove -frename-registers -freorder-blocks @gol
370 -freorder-blocks-and-partition -freorder-functions @gol
371 -frerun-cse-after-loop -freschedule-modulo-scheduled-loops @gol
372 -frounding-math -fsched2-use-superblocks -fsched-pressure @gol
373 -fsched-spec-load -fsched-spec-load-dangerous @gol
374 -fsched-stalled-insns-dep[=@var{n}] -fsched-stalled-insns[=@var{n}] @gol
375 -fsched-group-heuristic -fsched-critical-path-heuristic @gol
376 -fsched-spec-insn-heuristic -fsched-rank-heuristic @gol
377 -fsched-last-insn-heuristic -fsched-dep-count-heuristic @gol
378 -fschedule-insns -fschedule-insns2 -fsection-anchors @gol
379 -fselective-scheduling -fselective-scheduling2 @gol
380 -fsel-sched-pipelining -fsel-sched-pipelining-outer-loops @gol
381 -fsignaling-nans -fsingle-precision-constant -fsplit-ivs-in-unroller @gol
382 -fsplit-wide-types -fstack-protector -fstack-protector-all @gol
383 -fstrict-aliasing -fstrict-overflow -fthread-jumps -ftracer @gol
384 -ftree-builtin-call-dce -ftree-ccp -ftree-ch -ftree-copy-prop @gol
385 -ftree-copyrename -ftree-dce @gol
386 -ftree-dominator-opts -ftree-dse -ftree-forwprop -ftree-fre -ftree-loop-im @gol
387 -ftree-phiprop -ftree-loop-distribution @gol
388 -ftree-loop-ivcanon -ftree-loop-linear -ftree-loop-optimize @gol
389 -ftree-parallelize-loops=@var{n} -ftree-pre -ftree-pta -ftree-reassoc @gol
390 -ftree-sink -ftree-sra -ftree-switch-conversion @gol
391 -ftree-ter -ftree-vect-loop-version -ftree-vectorize -ftree-vrp @gol
392 -funit-at-a-time -funroll-all-loops -funroll-loops @gol
393 -funsafe-loop-optimizations -funsafe-math-optimizations -funswitch-loops @gol
394 -fvariable-expansion-in-unroller -fvect-cost-model -fvpt -fweb @gol
395 -fwhole-program -fwhopr[=@var{n}] -fwpa -fuse-linker-plugin @gol
396 --param @var{name}=@var{value}
397 -O -O0 -O1 -O2 -O3 -Os -Ofast}
399 @item Preprocessor Options
400 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
401 @gccoptlist{-A@var{question}=@var{answer} @gol
402 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
403 -C -dD -dI -dM -dN @gol
404 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
405 -idirafter @var{dir} @gol
406 -include @var{file} -imacros @var{file} @gol
407 -iprefix @var{file} -iwithprefix @var{dir} @gol
408 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
409 -imultilib @var{dir} -isysroot @var{dir} @gol
410 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
411 -P -fworking-directory -remap @gol
412 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
413 -Xpreprocessor @var{option}}
415 @item Assembler Option
416 @xref{Assembler Options,,Passing Options to the Assembler}.
417 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
420 @xref{Link Options,,Options for Linking}.
421 @gccoptlist{@var{object-file-name} -l@var{library} @gol
422 -nostartfiles -nodefaultlibs -nostdlib -pie -rdynamic @gol
423 -s -static -static-libgcc -static-libstdc++ -shared @gol
424 -shared-libgcc -symbolic @gol
425 -T @var{script} -Wl,@var{option} -Xlinker @var{option} @gol
428 @item Directory Options
429 @xref{Directory Options,,Options for Directory Search}.
430 @gccoptlist{-B@var{prefix} -I@var{dir} -iplugindir=@var{dir}}
431 -iquote@var{dir} -L@var{dir} -specs=@var{file} -I-
434 @item Machine Dependent Options
435 @xref{Submodel Options,,Hardware Models and Configurations}.
436 @c This list is ordered alphanumerically by subsection name.
437 @c Try and put the significant identifier (CPU or system) first,
438 @c so users have a clue at guessing where the ones they want will be.
441 @gccoptlist{-EB -EL @gol
442 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
443 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
446 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
447 -mabi=@var{name} @gol
448 -mapcs-stack-check -mno-apcs-stack-check @gol
449 -mapcs-float -mno-apcs-float @gol
450 -mapcs-reentrant -mno-apcs-reentrant @gol
451 -msched-prolog -mno-sched-prolog @gol
452 -mlittle-endian -mbig-endian -mwords-little-endian @gol
453 -mfloat-abi=@var{name} -msoft-float -mhard-float -mfpe @gol
454 -mfp16-format=@var{name}
455 -mthumb-interwork -mno-thumb-interwork @gol
456 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
457 -mstructure-size-boundary=@var{n} @gol
458 -mabort-on-noreturn @gol
459 -mlong-calls -mno-long-calls @gol
460 -msingle-pic-base -mno-single-pic-base @gol
461 -mpic-register=@var{reg} @gol
462 -mnop-fun-dllimport @gol
463 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
464 -mpoke-function-name @gol
466 -mtpcs-frame -mtpcs-leaf-frame @gol
467 -mcaller-super-interworking -mcallee-super-interworking @gol
469 -mword-relocations @gol
470 -mfix-cortex-m3-ldrd}
473 @gccoptlist{-mmcu=@var{mcu} -mno-interrupts @gol
474 -mcall-prologues -mtiny-stack -mint8}
476 @emph{Blackfin Options}
477 @gccoptlist{-mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]} @gol
478 -msim -momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol
479 -mspecld-anomaly -mno-specld-anomaly -mcsync-anomaly -mno-csync-anomaly @gol
480 -mlow-64k -mno-low64k -mstack-check-l1 -mid-shared-library @gol
481 -mno-id-shared-library -mshared-library-id=@var{n} @gol
482 -mleaf-id-shared-library -mno-leaf-id-shared-library @gol
483 -msep-data -mno-sep-data -mlong-calls -mno-long-calls @gol
484 -mfast-fp -minline-plt -mmulticore -mcorea -mcoreb -msdram @gol
488 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
489 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
490 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
491 -mstack-align -mdata-align -mconst-align @gol
492 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
493 -melf -maout -melinux -mlinux -sim -sim2 @gol
494 -mmul-bug-workaround -mno-mul-bug-workaround}
497 @gccoptlist{-mmac -mpush-args}
499 @emph{Darwin Options}
500 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
501 -arch_only -bind_at_load -bundle -bundle_loader @gol
502 -client_name -compatibility_version -current_version @gol
504 -dependency-file -dylib_file -dylinker_install_name @gol
505 -dynamic -dynamiclib -exported_symbols_list @gol
506 -filelist -flat_namespace -force_cpusubtype_ALL @gol
507 -force_flat_namespace -headerpad_max_install_names @gol
509 -image_base -init -install_name -keep_private_externs @gol
510 -multi_module -multiply_defined -multiply_defined_unused @gol
511 -noall_load -no_dead_strip_inits_and_terms @gol
512 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
513 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
514 -private_bundle -read_only_relocs -sectalign @gol
515 -sectobjectsymbols -whyload -seg1addr @gol
516 -sectcreate -sectobjectsymbols -sectorder @gol
517 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
518 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
519 -segprot -segs_read_only_addr -segs_read_write_addr @gol
520 -single_module -static -sub_library -sub_umbrella @gol
521 -twolevel_namespace -umbrella -undefined @gol
522 -unexported_symbols_list -weak_reference_mismatches @gol
523 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
524 -mkernel -mone-byte-bool}
526 @emph{DEC Alpha Options}
527 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
528 -mieee -mieee-with-inexact -mieee-conformant @gol
529 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
530 -mtrap-precision=@var{mode} -mbuild-constants @gol
531 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
532 -mbwx -mmax -mfix -mcix @gol
533 -mfloat-vax -mfloat-ieee @gol
534 -mexplicit-relocs -msmall-data -mlarge-data @gol
535 -msmall-text -mlarge-text @gol
536 -mmemory-latency=@var{time}}
538 @emph{DEC Alpha/VMS Options}
539 @gccoptlist{-mvms-return-codes -mdebug-main=@var{prefix} -mmalloc64}
542 @gccoptlist{-msmall-model -mno-lsim}
545 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
546 -mhard-float -msoft-float @gol
547 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
548 -mdouble -mno-double @gol
549 -mmedia -mno-media -mmuladd -mno-muladd @gol
550 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
551 -mlinked-fp -mlong-calls -malign-labels @gol
552 -mlibrary-pic -macc-4 -macc-8 @gol
553 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
554 -moptimize-membar -mno-optimize-membar @gol
555 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
556 -mvliw-branch -mno-vliw-branch @gol
557 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
558 -mno-nested-cond-exec -mtomcat-stats @gol
562 @emph{GNU/Linux Options}
563 @gccoptlist{-mglibc -muclibc -mbionic -mandroid @gol
564 -tno-android-cc -tno-android-ld}
566 @emph{H8/300 Options}
567 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
570 @gccoptlist{-march=@var{architecture-type} @gol
571 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
572 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
573 -mfixed-range=@var{register-range} @gol
574 -mjump-in-delay -mlinker-opt -mlong-calls @gol
575 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
576 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
577 -mno-jump-in-delay -mno-long-load-store @gol
578 -mno-portable-runtime -mno-soft-float @gol
579 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
580 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
581 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
582 -munix=@var{unix-std} -nolibdld -static -threads}
584 @emph{i386 and x86-64 Options}
585 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
586 -mfpmath=@var{unit} @gol
587 -masm=@var{dialect} -mno-fancy-math-387 @gol
588 -mno-fp-ret-in-387 -msoft-float @gol
589 -mno-wide-multiply -mrtd -malign-double @gol
590 -mpreferred-stack-boundary=@var{num}
591 -mincoming-stack-boundary=@var{num}
592 -mcld -mcx16 -msahf -mmovbe -mcrc32 -mrecip @gol
593 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4 -mavx @gol
594 -maes -mpclmul -mfused-madd @gol
595 -msse4a -m3dnow -mpopcnt -mabm -mfma4 -mxop -mlwp @gol
596 -mthreads -mno-align-stringops -minline-all-stringops @gol
597 -minline-stringops-dynamically -mstringop-strategy=@var{alg} @gol
598 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
599 -m96bit-long-double -mregparm=@var{num} -msseregparm @gol
600 -mveclibabi=@var{type} -mpc32 -mpc64 -mpc80 -mstackrealign @gol
601 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
602 -mcmodel=@var{code-model} -mabi=@var{name} @gol
603 -m32 -m64 -mlarge-data-threshold=@var{num} @gol
607 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
608 -mvolatile-asm-stop -mregister-names -msdata -mno-sdata @gol
609 -mconstant-gp -mauto-pic -mfused-madd @gol
610 -minline-float-divide-min-latency @gol
611 -minline-float-divide-max-throughput @gol
612 -mno-inline-float-divide @gol
613 -minline-int-divide-min-latency @gol
614 -minline-int-divide-max-throughput @gol
615 -mno-inline-int-divide @gol
616 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
617 -mno-inline-sqrt @gol
618 -mdwarf2-asm -mearly-stop-bits @gol
619 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
620 -mtune=@var{cpu-type} -milp32 -mlp64 @gol
621 -msched-br-data-spec -msched-ar-data-spec -msched-control-spec @gol
622 -msched-br-in-data-spec -msched-ar-in-data-spec -msched-in-control-spec @gol
623 -msched-spec-ldc -msched-spec-control-ldc @gol
624 -msched-prefer-non-data-spec-insns -msched-prefer-non-control-spec-insns @gol
625 -msched-stop-bits-after-every-cycle -msched-count-spec-in-critical-path @gol
626 -msel-sched-dont-check-control-spec -msched-fp-mem-deps-zero-cost @gol
627 -msched-max-memory-insns-hard-limit -msched-max-memory-insns=@var{max-insns}}
629 @emph{IA-64/VMS Options}
630 @gccoptlist{-mvms-return-codes -mdebug-main=@var{prefix} -mmalloc64}
633 @gccoptlist{-mbarrel-shift-enabled -mdivide-enabled -mmultiply-enabled @gol
634 -msign-extend-enabled -muser-enabled}
636 @emph{M32R/D Options}
637 @gccoptlist{-m32r2 -m32rx -m32r @gol
639 -malign-loops -mno-align-loops @gol
640 -missue-rate=@var{number} @gol
641 -mbranch-cost=@var{number} @gol
642 -mmodel=@var{code-size-model-type} @gol
643 -msdata=@var{sdata-type} @gol
644 -mno-flush-func -mflush-func=@var{name} @gol
645 -mno-flush-trap -mflush-trap=@var{number} @gol
649 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
651 @emph{M680x0 Options}
652 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} -mtune=@var{tune}
653 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
654 -m68060 -mcpu32 -m5200 -m5206e -m528x -m5307 -m5407 @gol
655 -mcfv4e -mbitfield -mno-bitfield -mc68000 -mc68020 @gol
656 -mnobitfield -mrtd -mno-rtd -mdiv -mno-div -mshort @gol
657 -mno-short -mhard-float -m68881 -msoft-float -mpcrel @gol
658 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
659 -mshared-library-id=n -mid-shared-library -mno-id-shared-library @gol
662 @emph{M68hc1x Options}
663 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
664 -mauto-incdec -minmax -mlong-calls -mshort @gol
665 -msoft-reg-count=@var{count}}
668 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
669 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
670 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
671 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
672 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
675 @gccoptlist{-mabsdiff -mall-opts -maverage -mbased=@var{n} -mbitops @gol
676 -mc=@var{n} -mclip -mconfig=@var{name} -mcop -mcop32 -mcop64 -mivc2 @gol
677 -mdc -mdiv -meb -mel -mio-volatile -ml -mleadz -mm -mminmax @gol
678 -mmult -mno-opts -mrepeat -ms -msatur -msdram -msim -msimnovec -mtf @gol
682 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
683 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 @gol
684 -mips64 -mips64r2 @gol
685 -mips16 -mno-mips16 -mflip-mips16 @gol
686 -minterlink-mips16 -mno-interlink-mips16 @gol
687 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
688 -mshared -mno-shared -mplt -mno-plt -mxgot -mno-xgot @gol
689 -mgp32 -mgp64 -mfp32 -mfp64 -mhard-float -msoft-float @gol
690 -msingle-float -mdouble-float -mdsp -mno-dsp -mdspr2 -mno-dspr2 @gol
691 -mfpu=@var{fpu-type} @gol
692 -msmartmips -mno-smartmips @gol
693 -mpaired-single -mno-paired-single -mdmx -mno-mdmx @gol
694 -mips3d -mno-mips3d -mmt -mno-mt -mllsc -mno-llsc @gol
695 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
696 -G@var{num} -mlocal-sdata -mno-local-sdata @gol
697 -mextern-sdata -mno-extern-sdata -mgpopt -mno-gopt @gol
698 -membedded-data -mno-embedded-data @gol
699 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
700 -mcode-readable=@var{setting} @gol
701 -msplit-addresses -mno-split-addresses @gol
702 -mexplicit-relocs -mno-explicit-relocs @gol
703 -mcheck-zero-division -mno-check-zero-division @gol
704 -mdivide-traps -mdivide-breaks @gol
705 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
706 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
707 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
708 -mfix-r10000 -mno-fix-r10000 -mfix-vr4120 -mno-fix-vr4120 @gol
709 -mfix-vr4130 -mno-fix-vr4130 -mfix-sb1 -mno-fix-sb1 @gol
710 -mflush-func=@var{func} -mno-flush-func @gol
711 -mbranch-cost=@var{num} -mbranch-likely -mno-branch-likely @gol
712 -mfp-exceptions -mno-fp-exceptions @gol
713 -mvr4130-align -mno-vr4130-align -msynci -mno-synci @gol
714 -mrelax-pic-calls -mno-relax-pic-calls -mmcount-ra-address}
717 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
718 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
719 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
720 -mno-base-addresses -msingle-exit -mno-single-exit}
722 @emph{MN10300 Options}
723 @gccoptlist{-mmult-bug -mno-mult-bug @gol
724 -mam33 -mno-am33 @gol
725 -mam33-2 -mno-am33-2 @gol
726 -mreturn-pointer-on-d0 @gol
729 @emph{PDP-11 Options}
730 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
731 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
732 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
733 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
734 -mbranch-expensive -mbranch-cheap @gol
735 -msplit -mno-split -munix-asm -mdec-asm}
737 @emph{picoChip Options}
738 @gccoptlist{-mae=@var{ae_type} -mvliw-lookahead=@var{N}
739 -msymbol-as-address -mno-inefficient-warnings}
741 @emph{PowerPC Options}
742 See RS/6000 and PowerPC Options.
744 @emph{RS/6000 and PowerPC Options}
745 @gccoptlist{-mcpu=@var{cpu-type} @gol
746 -mtune=@var{cpu-type} @gol
747 -mcmodel=@var{code-model} @gol
748 -mpower -mno-power -mpower2 -mno-power2 @gol
749 -mpowerpc -mpowerpc64 -mno-powerpc @gol
750 -maltivec -mno-altivec @gol
751 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
752 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
753 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mpopcntd -mno-popcntd @gol
754 -mfprnd -mno-fprnd @gol
755 -mcmpb -mno-cmpb -mmfpgpr -mno-mfpgpr -mhard-dfp -mno-hard-dfp @gol
756 -mnew-mnemonics -mold-mnemonics @gol
757 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
758 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
759 -malign-power -malign-natural @gol
760 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
761 -msingle-float -mdouble-float -msimple-fpu @gol
762 -mstring -mno-string -mupdate -mno-update @gol
763 -mavoid-indexed-addresses -mno-avoid-indexed-addresses @gol
764 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
765 -mstrict-align -mno-strict-align -mrelocatable @gol
766 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
767 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
768 -mdynamic-no-pic -maltivec -mswdiv @gol
769 -mprioritize-restricted-insns=@var{priority} @gol
770 -msched-costly-dep=@var{dependence_type} @gol
771 -minsert-sched-nops=@var{scheme} @gol
772 -mcall-sysv -mcall-netbsd @gol
773 -maix-struct-return -msvr4-struct-return @gol
774 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
775 -misel -mno-isel @gol
776 -misel=yes -misel=no @gol
778 -mspe=yes -mspe=no @gol
780 -mgen-cell-microcode -mwarn-cell-microcode @gol
781 -mvrsave -mno-vrsave @gol
782 -mmulhw -mno-mulhw @gol
783 -mdlmzb -mno-dlmzb @gol
784 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
785 -mprototype -mno-prototype @gol
786 -msim -mmvme -mads -myellowknife -memb -msdata @gol
787 -msdata=@var{opt} -mvxworks -G @var{num} -pthread @gol
788 -mrecip -mrecip=@var{opt} -mno-recip -mrecip-precision -mno-recip-precision}
791 @gccoptlist{-m64bit-doubles -m32bit-doubles -fpu -nofpu@gol
793 -mbig-endian-data -mlittle-endian-data @gol
796 -mas100-syntax -mno-as100-syntax@gol
798 -mmax-constant-size=@gol
800 -msave-acc-in-interrupts}
802 @emph{S/390 and zSeries Options}
803 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
804 -mhard-float -msoft-float -mhard-dfp -mno-hard-dfp @gol
805 -mlong-double-64 -mlong-double-128 @gol
806 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
807 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
808 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
809 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
810 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard}
813 @gccoptlist{-meb -mel @gol
817 -mscore5 -mscore5u -mscore7 -mscore7d}
820 @gccoptlist{-m1 -m2 -m2e @gol
821 -m2a-nofpu -m2a-single-only -m2a-single -m2a @gol
823 -m4-nofpu -m4-single-only -m4-single -m4 @gol
824 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
825 -m5-64media -m5-64media-nofpu @gol
826 -m5-32media -m5-32media-nofpu @gol
827 -m5-compact -m5-compact-nofpu @gol
828 -mb -ml -mdalign -mrelax @gol
829 -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol
830 -mieee -mbitops -misize -minline-ic_invalidate -mpadstruct -mspace @gol
831 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
832 -mdivsi3_libfunc=@var{name} -mfixed-range=@var{register-range} @gol
833 -madjust-unroll -mindexed-addressing -mgettrcost=@var{number} -mpt-fixed @gol
834 -maccumulate-outgoing-args -minvalid-symbols}
837 @gccoptlist{-mcpu=@var{cpu-type} @gol
838 -mtune=@var{cpu-type} @gol
839 -mcmodel=@var{code-model} @gol
840 -m32 -m64 -mapp-regs -mno-app-regs @gol
841 -mfaster-structs -mno-faster-structs @gol
842 -mfpu -mno-fpu -mhard-float -msoft-float @gol
843 -mhard-quad-float -msoft-quad-float @gol
844 -mimpure-text -mno-impure-text -mlittle-endian @gol
845 -mstack-bias -mno-stack-bias @gol
846 -munaligned-doubles -mno-unaligned-doubles @gol
847 -mv8plus -mno-v8plus -mvis -mno-vis
848 -threads -pthreads -pthread}
851 @gccoptlist{-mwarn-reloc -merror-reloc @gol
852 -msafe-dma -munsafe-dma @gol
854 -msmall-mem -mlarge-mem -mstdmain @gol
855 -mfixed-range=@var{register-range} @gol
857 -maddress-space-conversion -mno-address-space-conversion @gol
858 -mcache-size=@var{cache-size} @gol
859 -matomic-updates -mno-atomic-updates}
861 @emph{System V Options}
862 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
865 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
866 -mprolog-function -mno-prolog-function -mspace @gol
867 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
868 -mapp-regs -mno-app-regs @gol
869 -mdisable-callt -mno-disable-callt @gol
875 @gccoptlist{-mg -mgnu -munix}
877 @emph{VxWorks Options}
878 @gccoptlist{-mrtp -non-static -Bstatic -Bdynamic @gol
879 -Xbind-lazy -Xbind-now}
881 @emph{x86-64 Options}
882 See i386 and x86-64 Options.
884 @emph{i386 and x86-64 Windows Options}
885 @gccoptlist{-mconsole -mcygwin -mno-cygwin -mdll
886 -mnop-fun-dllimport -mthread -municode -mwin32 -mwindows
887 -fno-set-stack-executable}
889 @emph{Xstormy16 Options}
892 @emph{Xtensa Options}
893 @gccoptlist{-mconst16 -mno-const16 @gol
894 -mfused-madd -mno-fused-madd @gol
896 -mserialize-volatile -mno-serialize-volatile @gol
897 -mtext-section-literals -mno-text-section-literals @gol
898 -mtarget-align -mno-target-align @gol
899 -mlongcalls -mno-longcalls}
901 @emph{zSeries Options}
902 See S/390 and zSeries Options.
904 @item Code Generation Options
905 @xref{Code Gen Options,,Options for Code Generation Conventions}.
906 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
907 -ffixed-@var{reg} -fexceptions @gol
908 -fnon-call-exceptions -funwind-tables @gol
909 -fasynchronous-unwind-tables @gol
910 -finhibit-size-directive -finstrument-functions @gol
911 -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{} @gol
912 -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{} @gol
913 -fno-common -fno-ident @gol
914 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
915 -fno-jump-tables @gol
916 -frecord-gcc-switches @gol
917 -freg-struct-return -fshort-enums @gol
918 -fshort-double -fshort-wchar @gol
919 -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol
920 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
921 -fno-stack-limit @gol
922 -fleading-underscore -ftls-model=@var{model} @gol
923 -ftrapv -fwrapv -fbounds-check @gol
928 * Overall Options:: Controlling the kind of output:
929 an executable, object files, assembler files,
930 or preprocessed source.
931 * C Dialect Options:: Controlling the variant of C language compiled.
932 * C++ Dialect Options:: Variations on C++.
933 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
935 * Language Independent Options:: Controlling how diagnostics should be
937 * Warning Options:: How picky should the compiler be?
938 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
939 * Optimize Options:: How much optimization?
940 * Preprocessor Options:: Controlling header files and macro definitions.
941 Also, getting dependency information for Make.
942 * Assembler Options:: Passing options to the assembler.
943 * Link Options:: Specifying libraries and so on.
944 * Directory Options:: Where to find header files and libraries.
945 Where to find the compiler executable files.
946 * Spec Files:: How to pass switches to sub-processes.
947 * Target Options:: Running a cross-compiler, or an old version of GCC.
950 @node Overall Options
951 @section Options Controlling the Kind of Output
953 Compilation can involve up to four stages: preprocessing, compilation
954 proper, assembly and linking, always in that order. GCC is capable of
955 preprocessing and compiling several files either into several
956 assembler input files, or into one assembler input file; then each
957 assembler input file produces an object file, and linking combines all
958 the object files (those newly compiled, and those specified as input)
959 into an executable file.
961 @cindex file name suffix
962 For any given input file, the file name suffix determines what kind of
967 C source code which must be preprocessed.
970 C source code which should not be preprocessed.
973 C++ source code which should not be preprocessed.
976 Objective-C source code. Note that you must link with the @file{libobjc}
977 library to make an Objective-C program work.
980 Objective-C source code which should not be preprocessed.
984 Objective-C++ source code. Note that you must link with the @file{libobjc}
985 library to make an Objective-C++ program work. Note that @samp{.M} refers
986 to a literal capital M@.
989 Objective-C++ source code which should not be preprocessed.
992 C, C++, Objective-C or Objective-C++ header file to be turned into a
993 precompiled header (default), or C, C++ header file to be turned into an
994 Ada spec (via the @option{-fdump-ada-spec} switch).
998 @itemx @var{file}.cxx
999 @itemx @var{file}.cpp
1000 @itemx @var{file}.CPP
1001 @itemx @var{file}.c++
1003 C++ source code which must be preprocessed. Note that in @samp{.cxx},
1004 the last two letters must both be literally @samp{x}. Likewise,
1005 @samp{.C} refers to a literal capital C@.
1009 Objective-C++ source code which must be preprocessed.
1011 @item @var{file}.mii
1012 Objective-C++ source code which should not be preprocessed.
1016 @itemx @var{file}.hp
1017 @itemx @var{file}.hxx
1018 @itemx @var{file}.hpp
1019 @itemx @var{file}.HPP
1020 @itemx @var{file}.h++
1021 @itemx @var{file}.tcc
1022 C++ header file to be turned into a precompiled header or Ada spec.
1025 @itemx @var{file}.for
1026 @itemx @var{file}.ftn
1027 Fixed form Fortran source code which should not be preprocessed.
1030 @itemx @var{file}.FOR
1031 @itemx @var{file}.fpp
1032 @itemx @var{file}.FPP
1033 @itemx @var{file}.FTN
1034 Fixed form Fortran source code which must be preprocessed (with the traditional
1037 @item @var{file}.f90
1038 @itemx @var{file}.f95
1039 @itemx @var{file}.f03
1040 @itemx @var{file}.f08
1041 Free form Fortran source code which should not be preprocessed.
1043 @item @var{file}.F90
1044 @itemx @var{file}.F95
1045 @itemx @var{file}.F03
1046 @itemx @var{file}.F08
1047 Free form Fortran source code which must be preprocessed (with the
1048 traditional preprocessor).
1050 @c FIXME: Descriptions of Java file types.
1056 @item @var{file}.ads
1057 Ada source code file which contains a library unit declaration (a
1058 declaration of a package, subprogram, or generic, or a generic
1059 instantiation), or a library unit renaming declaration (a package,
1060 generic, or subprogram renaming declaration). Such files are also
1063 @item @var{file}.adb
1064 Ada source code file containing a library unit body (a subprogram or
1065 package body). Such files are also called @dfn{bodies}.
1067 @c GCC also knows about some suffixes for languages not yet included:
1078 @itemx @var{file}.sx
1079 Assembler code which must be preprocessed.
1082 An object file to be fed straight into linking.
1083 Any file name with no recognized suffix is treated this way.
1087 You can specify the input language explicitly with the @option{-x} option:
1090 @item -x @var{language}
1091 Specify explicitly the @var{language} for the following input files
1092 (rather than letting the compiler choose a default based on the file
1093 name suffix). This option applies to all following input files until
1094 the next @option{-x} option. Possible values for @var{language} are:
1096 c c-header c-cpp-output
1097 c++ c++-header c++-cpp-output
1098 objective-c objective-c-header objective-c-cpp-output
1099 objective-c++ objective-c++-header objective-c++-cpp-output
1100 assembler assembler-with-cpp
1102 f77 f77-cpp-input f95 f95-cpp-input
1107 Turn off any specification of a language, so that subsequent files are
1108 handled according to their file name suffixes (as they are if @option{-x}
1109 has not been used at all).
1111 @item -pass-exit-codes
1112 @opindex pass-exit-codes
1113 Normally the @command{gcc} program will exit with the code of 1 if any
1114 phase of the compiler returns a non-success return code. If you specify
1115 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
1116 numerically highest error produced by any phase that returned an error
1117 indication. The C, C++, and Fortran frontends return 4, if an internal
1118 compiler error is encountered.
1121 If you only want some of the stages of compilation, you can use
1122 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
1123 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
1124 @command{gcc} is to stop. Note that some combinations (for example,
1125 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
1130 Compile or assemble the source files, but do not link. The linking
1131 stage simply is not done. The ultimate output is in the form of an
1132 object file for each source file.
1134 By default, the object file name for a source file is made by replacing
1135 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
1137 Unrecognized input files, not requiring compilation or assembly, are
1142 Stop after the stage of compilation proper; do not assemble. The output
1143 is in the form of an assembler code file for each non-assembler input
1146 By default, the assembler file name for a source file is made by
1147 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
1149 Input files that don't require compilation are ignored.
1153 Stop after the preprocessing stage; do not run the compiler proper. The
1154 output is in the form of preprocessed source code, which is sent to the
1157 Input files which don't require preprocessing are ignored.
1159 @cindex output file option
1162 Place output in file @var{file}. This applies regardless to whatever
1163 sort of output is being produced, whether it be an executable file,
1164 an object file, an assembler file or preprocessed C code.
1166 If @option{-o} is not specified, the default is to put an executable
1167 file in @file{a.out}, the object file for
1168 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1169 assembler file in @file{@var{source}.s}, a precompiled header file in
1170 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1175 Print (on standard error output) the commands executed to run the stages
1176 of compilation. Also print the version number of the compiler driver
1177 program and of the preprocessor and the compiler proper.
1181 Like @option{-v} except the commands are not executed and arguments
1182 are quoted unless they contain only alphanumeric characters or @code{./-_}.
1183 This is useful for shell scripts to capture the driver-generated command lines.
1187 Use pipes rather than temporary files for communication between the
1188 various stages of compilation. This fails to work on some systems where
1189 the assembler is unable to read from a pipe; but the GNU assembler has
1194 If you are compiling multiple source files, this option tells the driver
1195 to pass all the source files to the compiler at once (for those
1196 languages for which the compiler can handle this). This will allow
1197 intermodule analysis (IMA) to be performed by the compiler. Currently the only
1198 language for which this is supported is C@. If you pass source files for
1199 multiple languages to the driver, using this option, the driver will invoke
1200 the compiler(s) that support IMA once each, passing each compiler all the
1201 source files appropriate for it. For those languages that do not support
1202 IMA this option will be ignored, and the compiler will be invoked once for
1203 each source file in that language. If you use this option in conjunction
1204 with @option{-save-temps}, the compiler will generate multiple
1206 (one for each source file), but only one (combined) @file{.o} or
1211 Print (on the standard output) a description of the command line options
1212 understood by @command{gcc}. If the @option{-v} option is also specified
1213 then @option{--help} will also be passed on to the various processes
1214 invoked by @command{gcc}, so that they can display the command line options
1215 they accept. If the @option{-Wextra} option has also been specified
1216 (prior to the @option{--help} option), then command line options which
1217 have no documentation associated with them will also be displayed.
1220 @opindex target-help
1221 Print (on the standard output) a description of target-specific command
1222 line options for each tool. For some targets extra target-specific
1223 information may also be printed.
1225 @item --help=@{@var{class}@r{|[}^@r{]}@var{qualifier}@}@r{[},@dots{}@r{]}
1226 Print (on the standard output) a description of the command line
1227 options understood by the compiler that fit into all specified classes
1228 and qualifiers. These are the supported classes:
1231 @item @samp{optimizers}
1232 This will display all of the optimization options supported by the
1235 @item @samp{warnings}
1236 This will display all of the options controlling warning messages
1237 produced by the compiler.
1240 This will display target-specific options. Unlike the
1241 @option{--target-help} option however, target-specific options of the
1242 linker and assembler will not be displayed. This is because those
1243 tools do not currently support the extended @option{--help=} syntax.
1246 This will display the values recognized by the @option{--param}
1249 @item @var{language}
1250 This will display the options supported for @var{language}, where
1251 @var{language} is the name of one of the languages supported in this
1255 This will display the options that are common to all languages.
1258 These are the supported qualifiers:
1261 @item @samp{undocumented}
1262 Display only those options which are undocumented.
1265 Display options which take an argument that appears after an equal
1266 sign in the same continuous piece of text, such as:
1267 @samp{--help=target}.
1269 @item @samp{separate}
1270 Display options which take an argument that appears as a separate word
1271 following the original option, such as: @samp{-o output-file}.
1274 Thus for example to display all the undocumented target-specific
1275 switches supported by the compiler the following can be used:
1278 --help=target,undocumented
1281 The sense of a qualifier can be inverted by prefixing it with the
1282 @samp{^} character, so for example to display all binary warning
1283 options (i.e., ones that are either on or off and that do not take an
1284 argument), which have a description the following can be used:
1287 --help=warnings,^joined,^undocumented
1290 The argument to @option{--help=} should not consist solely of inverted
1293 Combining several classes is possible, although this usually
1294 restricts the output by so much that there is nothing to display. One
1295 case where it does work however is when one of the classes is
1296 @var{target}. So for example to display all the target-specific
1297 optimization options the following can be used:
1300 --help=target,optimizers
1303 The @option{--help=} option can be repeated on the command line. Each
1304 successive use will display its requested class of options, skipping
1305 those that have already been displayed.
1307 If the @option{-Q} option appears on the command line before the
1308 @option{--help=} option, then the descriptive text displayed by
1309 @option{--help=} is changed. Instead of describing the displayed
1310 options, an indication is given as to whether the option is enabled,
1311 disabled or set to a specific value (assuming that the compiler
1312 knows this at the point where the @option{--help=} option is used).
1314 Here is a truncated example from the ARM port of @command{gcc}:
1317 % gcc -Q -mabi=2 --help=target -c
1318 The following options are target specific:
1320 -mabort-on-noreturn [disabled]
1324 The output is sensitive to the effects of previous command line
1325 options, so for example it is possible to find out which optimizations
1326 are enabled at @option{-O2} by using:
1329 -Q -O2 --help=optimizers
1332 Alternatively you can discover which binary optimizations are enabled
1333 by @option{-O3} by using:
1336 gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
1337 gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
1338 diff /tmp/O2-opts /tmp/O3-opts | grep enabled
1341 @item -no-canonical-prefixes
1342 @opindex no-canonical-prefixes
1343 Do not expand any symbolic links, resolve references to @samp{/../}
1344 or @samp{/./}, or make the path absolute when generating a relative
1349 Display the version number and copyrights of the invoked GCC@.
1353 Invoke all subcommands under a wrapper program. It takes a single
1354 comma separated list as an argument, which will be used to invoke
1358 gcc -c t.c -wrapper gdb,--args
1361 This will invoke all subprograms of gcc under "gdb --args",
1362 thus cc1 invocation will be "gdb --args cc1 ...".
1364 @item -fplugin=@var{name}.so
1365 Load the plugin code in file @var{name}.so, assumed to be a
1366 shared object to be dlopen'd by the compiler. The base name of
1367 the shared object file is used to identify the plugin for the
1368 purposes of argument parsing (See
1369 @option{-fplugin-arg-@var{name}-@var{key}=@var{value}} below).
1370 Each plugin should define the callback functions specified in the
1373 @item -fplugin-arg-@var{name}-@var{key}=@var{value}
1374 Define an argument called @var{key} with a value of @var{value}
1375 for the plugin called @var{name}.
1377 @item -fdump-ada-spec@r{[}-slim@r{]}
1378 For C and C++ source and include files, generate corresponding Ada
1379 specs. @xref{Generating Ada Bindings for C and C++ headers,,, gnat_ugn,
1380 GNAT User's Guide}, which provides detailed documentation on this feature.
1382 @include @value{srcdir}/../libiberty/at-file.texi
1386 @section Compiling C++ Programs
1388 @cindex suffixes for C++ source
1389 @cindex C++ source file suffixes
1390 C++ source files conventionally use one of the suffixes @samp{.C},
1391 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1392 @samp{.cxx}; C++ header files often use @samp{.hh}, @samp{.hpp},
1393 @samp{.H}, or (for shared template code) @samp{.tcc}; and
1394 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1395 files with these names and compiles them as C++ programs even if you
1396 call the compiler the same way as for compiling C programs (usually
1397 with the name @command{gcc}).
1401 However, the use of @command{gcc} does not add the C++ library.
1402 @command{g++} is a program that calls GCC and treats @samp{.c},
1403 @samp{.h} and @samp{.i} files as C++ source files instead of C source
1404 files unless @option{-x} is used, and automatically specifies linking
1405 against the C++ library. This program is also useful when
1406 precompiling a C header file with a @samp{.h} extension for use in C++
1407 compilations. On many systems, @command{g++} is also installed with
1408 the name @command{c++}.
1410 @cindex invoking @command{g++}
1411 When you compile C++ programs, you may specify many of the same
1412 command-line options that you use for compiling programs in any
1413 language; or command-line options meaningful for C and related
1414 languages; or options that are meaningful only for C++ programs.
1415 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1416 explanations of options for languages related to C@.
1417 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1418 explanations of options that are meaningful only for C++ programs.
1420 @node C Dialect Options
1421 @section Options Controlling C Dialect
1422 @cindex dialect options
1423 @cindex language dialect options
1424 @cindex options, dialect
1426 The following options control the dialect of C (or languages derived
1427 from C, such as C++, Objective-C and Objective-C++) that the compiler
1431 @cindex ANSI support
1435 In C mode, this is equivalent to @samp{-std=c90}. In C++ mode, it is
1436 equivalent to @samp{-std=c++98}.
1438 This turns off certain features of GCC that are incompatible with ISO
1439 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1440 such as the @code{asm} and @code{typeof} keywords, and
1441 predefined macros such as @code{unix} and @code{vax} that identify the
1442 type of system you are using. It also enables the undesirable and
1443 rarely used ISO trigraph feature. For the C compiler,
1444 it disables recognition of C++ style @samp{//} comments as well as
1445 the @code{inline} keyword.
1447 The alternate keywords @code{__asm__}, @code{__extension__},
1448 @code{__inline__} and @code{__typeof__} continue to work despite
1449 @option{-ansi}. You would not want to use them in an ISO C program, of
1450 course, but it is useful to put them in header files that might be included
1451 in compilations done with @option{-ansi}. Alternate predefined macros
1452 such as @code{__unix__} and @code{__vax__} are also available, with or
1453 without @option{-ansi}.
1455 The @option{-ansi} option does not cause non-ISO programs to be
1456 rejected gratuitously. For that, @option{-pedantic} is required in
1457 addition to @option{-ansi}. @xref{Warning Options}.
1459 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1460 option is used. Some header files may notice this macro and refrain
1461 from declaring certain functions or defining certain macros that the
1462 ISO standard doesn't call for; this is to avoid interfering with any
1463 programs that might use these names for other things.
1465 Functions that would normally be built in but do not have semantics
1466 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1467 functions when @option{-ansi} is used. @xref{Other Builtins,,Other
1468 built-in functions provided by GCC}, for details of the functions
1473 Determine the language standard. @xref{Standards,,Language Standards
1474 Supported by GCC}, for details of these standard versions. This option
1475 is currently only supported when compiling C or C++.
1477 The compiler can accept several base standards, such as @samp{c90} or
1478 @samp{c++98}, and GNU dialects of those standards, such as
1479 @samp{gnu90} or @samp{gnu++98}. By specifying a base standard, the
1480 compiler will accept all programs following that standard and those
1481 using GNU extensions that do not contradict it. For example,
1482 @samp{-std=c90} turns off certain features of GCC that are
1483 incompatible with ISO C90, such as the @code{asm} and @code{typeof}
1484 keywords, but not other GNU extensions that do not have a meaning in
1485 ISO C90, such as omitting the middle term of a @code{?:}
1486 expression. On the other hand, by specifying a GNU dialect of a
1487 standard, all features the compiler support are enabled, even when
1488 those features change the meaning of the base standard and some
1489 strict-conforming programs may be rejected. The particular standard
1490 is used by @option{-pedantic} to identify which features are GNU
1491 extensions given that version of the standard. For example
1492 @samp{-std=gnu90 -pedantic} would warn about C++ style @samp{//}
1493 comments, while @samp{-std=gnu99 -pedantic} would not.
1495 A value for this option must be provided; possible values are
1501 Support all ISO C90 programs (certain GNU extensions that conflict
1502 with ISO C90 are disabled). Same as @option{-ansi} for C code.
1504 @item iso9899:199409
1505 ISO C90 as modified in amendment 1.
1511 ISO C99. Note that this standard is not yet fully supported; see
1512 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1513 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1516 ISO C1X, the draft of the next revision of the ISO C standard.
1517 Support is limited and experimental and features enabled by this
1518 option may be changed or removed if changed in or removed from the
1523 GNU dialect of ISO C90 (including some C99 features). This
1524 is the default for C code.
1528 GNU dialect of ISO C99. When ISO C99 is fully implemented in GCC,
1529 this will become the default. The name @samp{gnu9x} is deprecated.
1532 GNU dialect of ISO C1X. Support is limited and experimental and
1533 features enabled by this option may be changed or removed if changed
1534 in or removed from the standard draft.
1537 The 1998 ISO C++ standard plus amendments. Same as @option{-ansi} for
1541 GNU dialect of @option{-std=c++98}. This is the default for
1545 The working draft of the upcoming ISO C++0x standard. This option
1546 enables experimental features that are likely to be included in
1547 C++0x. The working draft is constantly changing, and any feature that is
1548 enabled by this flag may be removed from future versions of GCC if it is
1549 not part of the C++0x standard.
1552 GNU dialect of @option{-std=c++0x}. This option enables
1553 experimental features that may be removed in future versions of GCC.
1556 @item -fgnu89-inline
1557 @opindex fgnu89-inline
1558 The option @option{-fgnu89-inline} tells GCC to use the traditional
1559 GNU semantics for @code{inline} functions when in C99 mode.
1560 @xref{Inline,,An Inline Function is As Fast As a Macro}. This option
1561 is accepted and ignored by GCC versions 4.1.3 up to but not including
1562 4.3. In GCC versions 4.3 and later it changes the behavior of GCC in
1563 C99 mode. Using this option is roughly equivalent to adding the
1564 @code{gnu_inline} function attribute to all inline functions
1565 (@pxref{Function Attributes}).
1567 The option @option{-fno-gnu89-inline} explicitly tells GCC to use the
1568 C99 semantics for @code{inline} when in C99 or gnu99 mode (i.e., it
1569 specifies the default behavior). This option was first supported in
1570 GCC 4.3. This option is not supported in @option{-std=c90} or
1571 @option{-std=gnu90} mode.
1573 The preprocessor macros @code{__GNUC_GNU_INLINE__} and
1574 @code{__GNUC_STDC_INLINE__} may be used to check which semantics are
1575 in effect for @code{inline} functions. @xref{Common Predefined
1576 Macros,,,cpp,The C Preprocessor}.
1578 @item -aux-info @var{filename}
1580 Output to the given filename prototyped declarations for all functions
1581 declared and/or defined in a translation unit, including those in header
1582 files. This option is silently ignored in any language other than C@.
1584 Besides declarations, the file indicates, in comments, the origin of
1585 each declaration (source file and line), whether the declaration was
1586 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1587 @samp{O} for old, respectively, in the first character after the line
1588 number and the colon), and whether it came from a declaration or a
1589 definition (@samp{C} or @samp{F}, respectively, in the following
1590 character). In the case of function definitions, a K&R-style list of
1591 arguments followed by their declarations is also provided, inside
1592 comments, after the declaration.
1596 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1597 keyword, so that code can use these words as identifiers. You can use
1598 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1599 instead. @option{-ansi} implies @option{-fno-asm}.
1601 In C++, this switch only affects the @code{typeof} keyword, since
1602 @code{asm} and @code{inline} are standard keywords. You may want to
1603 use the @option{-fno-gnu-keywords} flag instead, which has the same
1604 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1605 switch only affects the @code{asm} and @code{typeof} keywords, since
1606 @code{inline} is a standard keyword in ISO C99.
1609 @itemx -fno-builtin-@var{function}
1610 @opindex fno-builtin
1611 @cindex built-in functions
1612 Don't recognize built-in functions that do not begin with
1613 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1614 functions provided by GCC}, for details of the functions affected,
1615 including those which are not built-in functions when @option{-ansi} or
1616 @option{-std} options for strict ISO C conformance are used because they
1617 do not have an ISO standard meaning.
1619 GCC normally generates special code to handle certain built-in functions
1620 more efficiently; for instance, calls to @code{alloca} may become single
1621 instructions that adjust the stack directly, and calls to @code{memcpy}
1622 may become inline copy loops. The resulting code is often both smaller
1623 and faster, but since the function calls no longer appear as such, you
1624 cannot set a breakpoint on those calls, nor can you change the behavior
1625 of the functions by linking with a different library. In addition,
1626 when a function is recognized as a built-in function, GCC may use
1627 information about that function to warn about problems with calls to
1628 that function, or to generate more efficient code, even if the
1629 resulting code still contains calls to that function. For example,
1630 warnings are given with @option{-Wformat} for bad calls to
1631 @code{printf}, when @code{printf} is built in, and @code{strlen} is
1632 known not to modify global memory.
1634 With the @option{-fno-builtin-@var{function}} option
1635 only the built-in function @var{function} is
1636 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1637 function is named that is not built-in in this version of GCC, this
1638 option is ignored. There is no corresponding
1639 @option{-fbuiltin-@var{function}} option; if you wish to enable
1640 built-in functions selectively when using @option{-fno-builtin} or
1641 @option{-ffreestanding}, you may define macros such as:
1644 #define abs(n) __builtin_abs ((n))
1645 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1650 @cindex hosted environment
1652 Assert that compilation takes place in a hosted environment. This implies
1653 @option{-fbuiltin}. A hosted environment is one in which the
1654 entire standard library is available, and in which @code{main} has a return
1655 type of @code{int}. Examples are nearly everything except a kernel.
1656 This is equivalent to @option{-fno-freestanding}.
1658 @item -ffreestanding
1659 @opindex ffreestanding
1660 @cindex hosted environment
1662 Assert that compilation takes place in a freestanding environment. This
1663 implies @option{-fno-builtin}. A freestanding environment
1664 is one in which the standard library may not exist, and program startup may
1665 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1666 This is equivalent to @option{-fno-hosted}.
1668 @xref{Standards,,Language Standards Supported by GCC}, for details of
1669 freestanding and hosted environments.
1673 @cindex openmp parallel
1674 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
1675 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
1676 compiler generates parallel code according to the OpenMP Application
1677 Program Interface v3.0 @w{@uref{http://www.openmp.org/}}. This option
1678 implies @option{-pthread}, and thus is only supported on targets that
1679 have support for @option{-pthread}.
1681 @item -fms-extensions
1682 @opindex fms-extensions
1683 Accept some non-standard constructs used in Microsoft header files.
1685 Some cases of unnamed fields in structures and unions are only
1686 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
1687 fields within structs/unions}, for details.
1691 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1692 options for strict ISO C conformance) implies @option{-trigraphs}.
1694 @item -no-integrated-cpp
1695 @opindex no-integrated-cpp
1696 Performs a compilation in two passes: preprocessing and compiling. This
1697 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1698 @option{-B} option. The user supplied compilation step can then add in
1699 an additional preprocessing step after normal preprocessing but before
1700 compiling. The default is to use the integrated cpp (internal cpp)
1702 The semantics of this option will change if "cc1", "cc1plus", and
1703 "cc1obj" are merged.
1705 @cindex traditional C language
1706 @cindex C language, traditional
1708 @itemx -traditional-cpp
1709 @opindex traditional-cpp
1710 @opindex traditional
1711 Formerly, these options caused GCC to attempt to emulate a pre-standard
1712 C compiler. They are now only supported with the @option{-E} switch.
1713 The preprocessor continues to support a pre-standard mode. See the GNU
1714 CPP manual for details.
1716 @item -fcond-mismatch
1717 @opindex fcond-mismatch
1718 Allow conditional expressions with mismatched types in the second and
1719 third arguments. The value of such an expression is void. This option
1720 is not supported for C++.
1722 @item -flax-vector-conversions
1723 @opindex flax-vector-conversions
1724 Allow implicit conversions between vectors with differing numbers of
1725 elements and/or incompatible element types. This option should not be
1728 @item -funsigned-char
1729 @opindex funsigned-char
1730 Let the type @code{char} be unsigned, like @code{unsigned char}.
1732 Each kind of machine has a default for what @code{char} should
1733 be. It is either like @code{unsigned char} by default or like
1734 @code{signed char} by default.
1736 Ideally, a portable program should always use @code{signed char} or
1737 @code{unsigned char} when it depends on the signedness of an object.
1738 But many programs have been written to use plain @code{char} and
1739 expect it to be signed, or expect it to be unsigned, depending on the
1740 machines they were written for. This option, and its inverse, let you
1741 make such a program work with the opposite default.
1743 The type @code{char} is always a distinct type from each of
1744 @code{signed char} or @code{unsigned char}, even though its behavior
1745 is always just like one of those two.
1748 @opindex fsigned-char
1749 Let the type @code{char} be signed, like @code{signed char}.
1751 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1752 the negative form of @option{-funsigned-char}. Likewise, the option
1753 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1755 @item -fsigned-bitfields
1756 @itemx -funsigned-bitfields
1757 @itemx -fno-signed-bitfields
1758 @itemx -fno-unsigned-bitfields
1759 @opindex fsigned-bitfields
1760 @opindex funsigned-bitfields
1761 @opindex fno-signed-bitfields
1762 @opindex fno-unsigned-bitfields
1763 These options control whether a bit-field is signed or unsigned, when the
1764 declaration does not use either @code{signed} or @code{unsigned}. By
1765 default, such a bit-field is signed, because this is consistent: the
1766 basic integer types such as @code{int} are signed types.
1769 @node C++ Dialect Options
1770 @section Options Controlling C++ Dialect
1772 @cindex compiler options, C++
1773 @cindex C++ options, command line
1774 @cindex options, C++
1775 This section describes the command-line options that are only meaningful
1776 for C++ programs; but you can also use most of the GNU compiler options
1777 regardless of what language your program is in. For example, you
1778 might compile a file @code{firstClass.C} like this:
1781 g++ -g -frepo -O -c firstClass.C
1785 In this example, only @option{-frepo} is an option meant
1786 only for C++ programs; you can use the other options with any
1787 language supported by GCC@.
1789 Here is a list of options that are @emph{only} for compiling C++ programs:
1793 @item -fabi-version=@var{n}
1794 @opindex fabi-version
1795 Use version @var{n} of the C++ ABI@. Version 2 is the version of the
1796 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1797 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1798 the version that conforms most closely to the C++ ABI specification.
1799 Therefore, the ABI obtained using version 0 will change as ABI bugs
1802 The default is version 2.
1804 Version 3 corrects an error in mangling a constant address as a
1807 Version 4 implements a standard mangling for vector types.
1809 See also @option{-Wabi}.
1811 @item -fno-access-control
1812 @opindex fno-access-control
1813 Turn off all access checking. This switch is mainly useful for working
1814 around bugs in the access control code.
1818 Check that the pointer returned by @code{operator new} is non-null
1819 before attempting to modify the storage allocated. This check is
1820 normally unnecessary because the C++ standard specifies that
1821 @code{operator new} will only return @code{0} if it is declared
1822 @samp{throw()}, in which case the compiler will always check the
1823 return value even without this option. In all other cases, when
1824 @code{operator new} has a non-empty exception specification, memory
1825 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1826 @samp{new (nothrow)}.
1828 @item -fconserve-space
1829 @opindex fconserve-space
1830 Put uninitialized or runtime-initialized global variables into the
1831 common segment, as C does. This saves space in the executable at the
1832 cost of not diagnosing duplicate definitions. If you compile with this
1833 flag and your program mysteriously crashes after @code{main()} has
1834 completed, you may have an object that is being destroyed twice because
1835 two definitions were merged.
1837 This option is no longer useful on most targets, now that support has
1838 been added for putting variables into BSS without making them common.
1840 @item -fno-deduce-init-list
1841 @opindex fno-deduce-init-list
1842 Disable deduction of a template type parameter as
1843 std::initializer_list from a brace-enclosed initializer list, i.e.
1846 template <class T> auto forward(T t) -> decltype (realfn (t))
1853 forward(@{1,2@}); // call forward<std::initializer_list<int>>
1857 This option is present because this deduction is an extension to the
1858 current specification in the C++0x working draft, and there was
1859 some concern about potential overload resolution problems.
1861 @item -ffriend-injection
1862 @opindex ffriend-injection
1863 Inject friend functions into the enclosing namespace, so that they are
1864 visible outside the scope of the class in which they are declared.
1865 Friend functions were documented to work this way in the old Annotated
1866 C++ Reference Manual, and versions of G++ before 4.1 always worked
1867 that way. However, in ISO C++ a friend function which is not declared
1868 in an enclosing scope can only be found using argument dependent
1869 lookup. This option causes friends to be injected as they were in
1872 This option is for compatibility, and may be removed in a future
1875 @item -fno-elide-constructors
1876 @opindex fno-elide-constructors
1877 The C++ standard allows an implementation to omit creating a temporary
1878 which is only used to initialize another object of the same type.
1879 Specifying this option disables that optimization, and forces G++ to
1880 call the copy constructor in all cases.
1882 @item -fno-enforce-eh-specs
1883 @opindex fno-enforce-eh-specs
1884 Don't generate code to check for violation of exception specifications
1885 at runtime. This option violates the C++ standard, but may be useful
1886 for reducing code size in production builds, much like defining
1887 @samp{NDEBUG}. This does not give user code permission to throw
1888 exceptions in violation of the exception specifications; the compiler
1889 will still optimize based on the specifications, so throwing an
1890 unexpected exception will result in undefined behavior.
1893 @itemx -fno-for-scope
1895 @opindex fno-for-scope
1896 If @option{-ffor-scope} is specified, the scope of variables declared in
1897 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1898 as specified by the C++ standard.
1899 If @option{-fno-for-scope} is specified, the scope of variables declared in
1900 a @i{for-init-statement} extends to the end of the enclosing scope,
1901 as was the case in old versions of G++, and other (traditional)
1902 implementations of C++.
1904 The default if neither flag is given to follow the standard,
1905 but to allow and give a warning for old-style code that would
1906 otherwise be invalid, or have different behavior.
1908 @item -fno-gnu-keywords
1909 @opindex fno-gnu-keywords
1910 Do not recognize @code{typeof} as a keyword, so that code can use this
1911 word as an identifier. You can use the keyword @code{__typeof__} instead.
1912 @option{-ansi} implies @option{-fno-gnu-keywords}.
1914 @item -fno-implicit-templates
1915 @opindex fno-implicit-templates
1916 Never emit code for non-inline templates which are instantiated
1917 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1918 @xref{Template Instantiation}, for more information.
1920 @item -fno-implicit-inline-templates
1921 @opindex fno-implicit-inline-templates
1922 Don't emit code for implicit instantiations of inline templates, either.
1923 The default is to handle inlines differently so that compiles with and
1924 without optimization will need the same set of explicit instantiations.
1926 @item -fno-implement-inlines
1927 @opindex fno-implement-inlines
1928 To save space, do not emit out-of-line copies of inline functions
1929 controlled by @samp{#pragma implementation}. This will cause linker
1930 errors if these functions are not inlined everywhere they are called.
1932 @item -fms-extensions
1933 @opindex fms-extensions
1934 Disable pedantic warnings about constructs used in MFC, such as implicit
1935 int and getting a pointer to member function via non-standard syntax.
1937 @item -fno-nonansi-builtins
1938 @opindex fno-nonansi-builtins
1939 Disable built-in declarations of functions that are not mandated by
1940 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1941 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1944 @opindex fnothrow-opt
1945 Treat a @code{throw()} exception specification as though it were a
1946 @code{noexcept} specification to reduce or eliminate the text size
1947 overhead relative to a function with no exception specification. If
1948 the function has local variables of types with non-trivial
1949 destructors, the exception specification will actually make the
1950 function smaller because the EH cleanups for those variables can be
1951 optimized away. The semantic effect is that an exception thrown out of
1952 a function with such an exception specification will result in a call
1953 to @code{terminate} rather than @code{unexpected}.
1955 @item -fno-operator-names
1956 @opindex fno-operator-names
1957 Do not treat the operator name keywords @code{and}, @code{bitand},
1958 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1959 synonyms as keywords.
1961 @item -fno-optional-diags
1962 @opindex fno-optional-diags
1963 Disable diagnostics that the standard says a compiler does not need to
1964 issue. Currently, the only such diagnostic issued by G++ is the one for
1965 a name having multiple meanings within a class.
1968 @opindex fpermissive
1969 Downgrade some diagnostics about nonconformant code from errors to
1970 warnings. Thus, using @option{-fpermissive} will allow some
1971 nonconforming code to compile.
1973 @item -fno-pretty-templates
1974 @opindex fno-pretty-templates
1975 When an error message refers to a specialization of a function
1976 template, the compiler will normally print the signature of the
1977 template followed by the template arguments and any typedefs or
1978 typenames in the signature (e.g. @code{void f(T) [with T = int]}
1979 rather than @code{void f(int)}) so that it's clear which template is
1980 involved. When an error message refers to a specialization of a class
1981 template, the compiler will omit any template arguments which match
1982 the default template arguments for that template. If either of these
1983 behaviors make it harder to understand the error message rather than
1984 easier, using @option{-fno-pretty-templates} will disable them.
1988 Enable automatic template instantiation at link time. This option also
1989 implies @option{-fno-implicit-templates}. @xref{Template
1990 Instantiation}, for more information.
1994 Disable generation of information about every class with virtual
1995 functions for use by the C++ runtime type identification features
1996 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1997 of the language, you can save some space by using this flag. Note that
1998 exception handling uses the same information, but it will generate it as
1999 needed. The @samp{dynamic_cast} operator can still be used for casts that
2000 do not require runtime type information, i.e.@: casts to @code{void *} or to
2001 unambiguous base classes.
2005 Emit statistics about front-end processing at the end of the compilation.
2006 This information is generally only useful to the G++ development team.
2008 @item -fstrict-enums
2009 @opindex fstrict-enums
2010 Allow the compiler to optimize using the assumption that a value of
2011 enumeration type can only be one of the values of the enumeration (as
2012 defined in the C++ standard; basically, a value which can be
2013 represented in the minimum number of bits needed to represent all the
2014 enumerators). This assumption may not be valid if the program uses a
2015 cast to convert an arbitrary integer value to the enumeration type.
2017 @item -ftemplate-depth=@var{n}
2018 @opindex ftemplate-depth
2019 Set the maximum instantiation depth for template classes to @var{n}.
2020 A limit on the template instantiation depth is needed to detect
2021 endless recursions during template class instantiation. ANSI/ISO C++
2022 conforming programs must not rely on a maximum depth greater than 17
2023 (changed to 1024 in C++0x).
2025 @item -fno-threadsafe-statics
2026 @opindex fno-threadsafe-statics
2027 Do not emit the extra code to use the routines specified in the C++
2028 ABI for thread-safe initialization of local statics. You can use this
2029 option to reduce code size slightly in code that doesn't need to be
2032 @item -fuse-cxa-atexit
2033 @opindex fuse-cxa-atexit
2034 Register destructors for objects with static storage duration with the
2035 @code{__cxa_atexit} function rather than the @code{atexit} function.
2036 This option is required for fully standards-compliant handling of static
2037 destructors, but will only work if your C library supports
2038 @code{__cxa_atexit}.
2040 @item -fno-use-cxa-get-exception-ptr
2041 @opindex fno-use-cxa-get-exception-ptr
2042 Don't use the @code{__cxa_get_exception_ptr} runtime routine. This
2043 will cause @code{std::uncaught_exception} to be incorrect, but is necessary
2044 if the runtime routine is not available.
2046 @item -fvisibility-inlines-hidden
2047 @opindex fvisibility-inlines-hidden
2048 This switch declares that the user does not attempt to compare
2049 pointers to inline methods where the addresses of the two functions
2050 were taken in different shared objects.
2052 The effect of this is that GCC may, effectively, mark inline methods with
2053 @code{__attribute__ ((visibility ("hidden")))} so that they do not
2054 appear in the export table of a DSO and do not require a PLT indirection
2055 when used within the DSO@. Enabling this option can have a dramatic effect
2056 on load and link times of a DSO as it massively reduces the size of the
2057 dynamic export table when the library makes heavy use of templates.
2059 The behavior of this switch is not quite the same as marking the
2060 methods as hidden directly, because it does not affect static variables
2061 local to the function or cause the compiler to deduce that
2062 the function is defined in only one shared object.
2064 You may mark a method as having a visibility explicitly to negate the
2065 effect of the switch for that method. For example, if you do want to
2066 compare pointers to a particular inline method, you might mark it as
2067 having default visibility. Marking the enclosing class with explicit
2068 visibility will have no effect.
2070 Explicitly instantiated inline methods are unaffected by this option
2071 as their linkage might otherwise cross a shared library boundary.
2072 @xref{Template Instantiation}.
2074 @item -fvisibility-ms-compat
2075 @opindex fvisibility-ms-compat
2076 This flag attempts to use visibility settings to make GCC's C++
2077 linkage model compatible with that of Microsoft Visual Studio.
2079 The flag makes these changes to GCC's linkage model:
2083 It sets the default visibility to @code{hidden}, like
2084 @option{-fvisibility=hidden}.
2087 Types, but not their members, are not hidden by default.
2090 The One Definition Rule is relaxed for types without explicit
2091 visibility specifications which are defined in more than one different
2092 shared object: those declarations are permitted if they would have
2093 been permitted when this option was not used.
2096 In new code it is better to use @option{-fvisibility=hidden} and
2097 export those classes which are intended to be externally visible.
2098 Unfortunately it is possible for code to rely, perhaps accidentally,
2099 on the Visual Studio behavior.
2101 Among the consequences of these changes are that static data members
2102 of the same type with the same name but defined in different shared
2103 objects will be different, so changing one will not change the other;
2104 and that pointers to function members defined in different shared
2105 objects may not compare equal. When this flag is given, it is a
2106 violation of the ODR to define types with the same name differently.
2110 Do not use weak symbol support, even if it is provided by the linker.
2111 By default, G++ will use weak symbols if they are available. This
2112 option exists only for testing, and should not be used by end-users;
2113 it will result in inferior code and has no benefits. This option may
2114 be removed in a future release of G++.
2118 Do not search for header files in the standard directories specific to
2119 C++, but do still search the other standard directories. (This option
2120 is used when building the C++ library.)
2123 In addition, these optimization, warning, and code generation options
2124 have meanings only for C++ programs:
2127 @item -fno-default-inline
2128 @opindex fno-default-inline
2129 Do not assume @samp{inline} for functions defined inside a class scope.
2130 @xref{Optimize Options,,Options That Control Optimization}. Note that these
2131 functions will have linkage like inline functions; they just won't be
2134 @item -Wabi @r{(C, Objective-C, C++ and Objective-C++ only)}
2137 Warn when G++ generates code that is probably not compatible with the
2138 vendor-neutral C++ ABI@. Although an effort has been made to warn about
2139 all such cases, there are probably some cases that are not warned about,
2140 even though G++ is generating incompatible code. There may also be
2141 cases where warnings are emitted even though the code that is generated
2144 You should rewrite your code to avoid these warnings if you are
2145 concerned about the fact that code generated by G++ may not be binary
2146 compatible with code generated by other compilers.
2148 The known incompatibilities in @option{-fabi-version=2} (the default) include:
2153 A template with a non-type template parameter of reference type is
2154 mangled incorrectly:
2157 template <int &> struct S @{@};
2161 This is fixed in @option{-fabi-version=3}.
2164 SIMD vector types declared using @code{__attribute ((vector_size))} are
2165 mangled in a non-standard way that does not allow for overloading of
2166 functions taking vectors of different sizes.
2168 The mangling is changed in @option{-fabi-version=4}.
2171 The known incompatibilities in @option{-fabi-version=1} include:
2176 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
2177 pack data into the same byte as a base class. For example:
2180 struct A @{ virtual void f(); int f1 : 1; @};
2181 struct B : public A @{ int f2 : 1; @};
2185 In this case, G++ will place @code{B::f2} into the same byte
2186 as@code{A::f1}; other compilers will not. You can avoid this problem
2187 by explicitly padding @code{A} so that its size is a multiple of the
2188 byte size on your platform; that will cause G++ and other compilers to
2189 layout @code{B} identically.
2192 Incorrect handling of tail-padding for virtual bases. G++ does not use
2193 tail padding when laying out virtual bases. For example:
2196 struct A @{ virtual void f(); char c1; @};
2197 struct B @{ B(); char c2; @};
2198 struct C : public A, public virtual B @{@};
2202 In this case, G++ will not place @code{B} into the tail-padding for
2203 @code{A}; other compilers will. You can avoid this problem by
2204 explicitly padding @code{A} so that its size is a multiple of its
2205 alignment (ignoring virtual base classes); that will cause G++ and other
2206 compilers to layout @code{C} identically.
2209 Incorrect handling of bit-fields with declared widths greater than that
2210 of their underlying types, when the bit-fields appear in a union. For
2214 union U @{ int i : 4096; @};
2218 Assuming that an @code{int} does not have 4096 bits, G++ will make the
2219 union too small by the number of bits in an @code{int}.
2222 Empty classes can be placed at incorrect offsets. For example:
2232 struct C : public B, public A @{@};
2236 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
2237 it should be placed at offset zero. G++ mistakenly believes that the
2238 @code{A} data member of @code{B} is already at offset zero.
2241 Names of template functions whose types involve @code{typename} or
2242 template template parameters can be mangled incorrectly.
2245 template <typename Q>
2246 void f(typename Q::X) @{@}
2248 template <template <typename> class Q>
2249 void f(typename Q<int>::X) @{@}
2253 Instantiations of these templates may be mangled incorrectly.
2257 It also warns psABI related changes. The known psABI changes at this
2263 For SYSV/x86-64, when passing union with long double, it is changed to
2264 pass in memory as specified in psABI. For example:
2274 @code{union U} will always be passed in memory.
2278 @item -Wctor-dtor-privacy @r{(C++ and Objective-C++ only)}
2279 @opindex Wctor-dtor-privacy
2280 @opindex Wno-ctor-dtor-privacy
2281 Warn when a class seems unusable because all the constructors or
2282 destructors in that class are private, and it has neither friends nor
2283 public static member functions.
2285 @item -Wnoexcept @r{(C++ and Objective-C++ only)}
2287 @opindex Wno-noexcept
2288 Warn when a noexcept-expression evaluates to false because of a call
2289 to a function that does not have a non-throwing exception
2290 specification (i.e. @samp{throw()} or @samp{noexcept}) but is known by
2291 the compiler to never throw an exception.
2293 @item -Wnon-virtual-dtor @r{(C++ and Objective-C++ only)}
2294 @opindex Wnon-virtual-dtor
2295 @opindex Wno-non-virtual-dtor
2296 Warn when a class has virtual functions and accessible non-virtual
2297 destructor, in which case it would be possible but unsafe to delete
2298 an instance of a derived class through a pointer to the base class.
2299 This warning is also enabled if -Weffc++ is specified.
2301 @item -Wreorder @r{(C++ and Objective-C++ only)}
2303 @opindex Wno-reorder
2304 @cindex reordering, warning
2305 @cindex warning for reordering of member initializers
2306 Warn when the order of member initializers given in the code does not
2307 match the order in which they must be executed. For instance:
2313 A(): j (0), i (1) @{ @}
2317 The compiler will rearrange the member initializers for @samp{i}
2318 and @samp{j} to match the declaration order of the members, emitting
2319 a warning to that effect. This warning is enabled by @option{-Wall}.
2322 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
2325 @item -Weffc++ @r{(C++ and Objective-C++ only)}
2328 Warn about violations of the following style guidelines from Scott Meyers'
2329 @cite{Effective C++} book:
2333 Item 11: Define a copy constructor and an assignment operator for classes
2334 with dynamically allocated memory.
2337 Item 12: Prefer initialization to assignment in constructors.
2340 Item 14: Make destructors virtual in base classes.
2343 Item 15: Have @code{operator=} return a reference to @code{*this}.
2346 Item 23: Don't try to return a reference when you must return an object.
2350 Also warn about violations of the following style guidelines from
2351 Scott Meyers' @cite{More Effective C++} book:
2355 Item 6: Distinguish between prefix and postfix forms of increment and
2356 decrement operators.
2359 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
2363 When selecting this option, be aware that the standard library
2364 headers do not obey all of these guidelines; use @samp{grep -v}
2365 to filter out those warnings.
2367 @item -Wstrict-null-sentinel @r{(C++ and Objective-C++ only)}
2368 @opindex Wstrict-null-sentinel
2369 @opindex Wno-strict-null-sentinel
2370 Warn also about the use of an uncasted @code{NULL} as sentinel. When
2371 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
2372 to @code{__null}. Although it is a null pointer constant not a null pointer,
2373 it is guaranteed to be of the same size as a pointer. But this use is
2374 not portable across different compilers.
2376 @item -Wno-non-template-friend @r{(C++ and Objective-C++ only)}
2377 @opindex Wno-non-template-friend
2378 @opindex Wnon-template-friend
2379 Disable warnings when non-templatized friend functions are declared
2380 within a template. Since the advent of explicit template specification
2381 support in G++, if the name of the friend is an unqualified-id (i.e.,
2382 @samp{friend foo(int)}), the C++ language specification demands that the
2383 friend declare or define an ordinary, nontemplate function. (Section
2384 14.5.3). Before G++ implemented explicit specification, unqualified-ids
2385 could be interpreted as a particular specialization of a templatized
2386 function. Because this non-conforming behavior is no longer the default
2387 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
2388 check existing code for potential trouble spots and is on by default.
2389 This new compiler behavior can be turned off with
2390 @option{-Wno-non-template-friend} which keeps the conformant compiler code
2391 but disables the helpful warning.
2393 @item -Wold-style-cast @r{(C++ and Objective-C++ only)}
2394 @opindex Wold-style-cast
2395 @opindex Wno-old-style-cast
2396 Warn if an old-style (C-style) cast to a non-void type is used within
2397 a C++ program. The new-style casts (@samp{dynamic_cast},
2398 @samp{static_cast}, @samp{reinterpret_cast}, and @samp{const_cast}) are
2399 less vulnerable to unintended effects and much easier to search for.
2401 @item -Woverloaded-virtual @r{(C++ and Objective-C++ only)}
2402 @opindex Woverloaded-virtual
2403 @opindex Wno-overloaded-virtual
2404 @cindex overloaded virtual fn, warning
2405 @cindex warning for overloaded virtual fn
2406 Warn when a function declaration hides virtual functions from a
2407 base class. For example, in:
2414 struct B: public A @{
2419 the @code{A} class version of @code{f} is hidden in @code{B}, and code
2427 will fail to compile.
2429 @item -Wno-pmf-conversions @r{(C++ and Objective-C++ only)}
2430 @opindex Wno-pmf-conversions
2431 @opindex Wpmf-conversions
2432 Disable the diagnostic for converting a bound pointer to member function
2435 @item -Wsign-promo @r{(C++ and Objective-C++ only)}
2436 @opindex Wsign-promo
2437 @opindex Wno-sign-promo
2438 Warn when overload resolution chooses a promotion from unsigned or
2439 enumerated type to a signed type, over a conversion to an unsigned type of
2440 the same size. Previous versions of G++ would try to preserve
2441 unsignedness, but the standard mandates the current behavior.
2446 A& operator = (int);
2456 In this example, G++ will synthesize a default @samp{A& operator =
2457 (const A&);}, while cfront will use the user-defined @samp{operator =}.
2460 @node Objective-C and Objective-C++ Dialect Options
2461 @section Options Controlling Objective-C and Objective-C++ Dialects
2463 @cindex compiler options, Objective-C and Objective-C++
2464 @cindex Objective-C and Objective-C++ options, command line
2465 @cindex options, Objective-C and Objective-C++
2466 (NOTE: This manual does not describe the Objective-C and Objective-C++
2467 languages themselves. See @xref{Standards,,Language Standards
2468 Supported by GCC}, for references.)
2470 This section describes the command-line options that are only meaningful
2471 for Objective-C and Objective-C++ programs, but you can also use most of
2472 the language-independent GNU compiler options.
2473 For example, you might compile a file @code{some_class.m} like this:
2476 gcc -g -fgnu-runtime -O -c some_class.m
2480 In this example, @option{-fgnu-runtime} is an option meant only for
2481 Objective-C and Objective-C++ programs; you can use the other options with
2482 any language supported by GCC@.
2484 Note that since Objective-C is an extension of the C language, Objective-C
2485 compilations may also use options specific to the C front-end (e.g.,
2486 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
2487 C++-specific options (e.g., @option{-Wabi}).
2489 Here is a list of options that are @emph{only} for compiling Objective-C
2490 and Objective-C++ programs:
2493 @item -fconstant-string-class=@var{class-name}
2494 @opindex fconstant-string-class
2495 Use @var{class-name} as the name of the class to instantiate for each
2496 literal string specified with the syntax @code{@@"@dots{}"}. The default
2497 class name is @code{NXConstantString} if the GNU runtime is being used, and
2498 @code{NSConstantString} if the NeXT runtime is being used (see below). The
2499 @option{-fconstant-cfstrings} option, if also present, will override the
2500 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
2501 to be laid out as constant CoreFoundation strings.
2504 @opindex fgnu-runtime
2505 Generate object code compatible with the standard GNU Objective-C
2506 runtime. This is the default for most types of systems.
2508 @item -fnext-runtime
2509 @opindex fnext-runtime
2510 Generate output compatible with the NeXT runtime. This is the default
2511 for NeXT-based systems, including Darwin and Mac OS X@. The macro
2512 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
2515 @item -fno-nil-receivers
2516 @opindex fno-nil-receivers
2517 Assume that all Objective-C message dispatches (e.g.,
2518 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
2519 is not @code{nil}. This allows for more efficient entry points in the runtime
2520 to be used. Currently, this option is only available in conjunction with
2521 the NeXT runtime on Mac OS X 10.3 and later.
2523 @item -fobjc-call-cxx-cdtors
2524 @opindex fobjc-call-cxx-cdtors
2525 For each Objective-C class, check if any of its instance variables is a
2526 C++ object with a non-trivial default constructor. If so, synthesize a
2527 special @code{- (id) .cxx_construct} instance method that will run
2528 non-trivial default constructors on any such instance variables, in order,
2529 and then return @code{self}. Similarly, check if any instance variable
2530 is a C++ object with a non-trivial destructor, and if so, synthesize a
2531 special @code{- (void) .cxx_destruct} method that will run
2532 all such default destructors, in reverse order.
2534 The @code{- (id) .cxx_construct} and/or @code{- (void) .cxx_destruct} methods
2535 thusly generated will only operate on instance variables declared in the
2536 current Objective-C class, and not those inherited from superclasses. It
2537 is the responsibility of the Objective-C runtime to invoke all such methods
2538 in an object's inheritance hierarchy. The @code{- (id) .cxx_construct} methods
2539 will be invoked by the runtime immediately after a new object
2540 instance is allocated; the @code{- (void) .cxx_destruct} methods will
2541 be invoked immediately before the runtime deallocates an object instance.
2543 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
2544 support for invoking the @code{- (id) .cxx_construct} and
2545 @code{- (void) .cxx_destruct} methods.
2547 @item -fobjc-direct-dispatch
2548 @opindex fobjc-direct-dispatch
2549 Allow fast jumps to the message dispatcher. On Darwin this is
2550 accomplished via the comm page.
2552 @item -fobjc-exceptions
2553 @opindex fobjc-exceptions
2554 Enable syntactic support for structured exception handling in Objective-C,
2555 similar to what is offered by C++ and Java. This option is
2556 unavailable in conjunction with the NeXT runtime on Mac OS X 10.2 and
2565 @@catch (AnObjCClass *exc) @{
2572 @@catch (AnotherClass *exc) @{
2575 @@catch (id allOthers) @{
2585 The @code{@@throw} statement may appear anywhere in an Objective-C or
2586 Objective-C++ program; when used inside of a @code{@@catch} block, the
2587 @code{@@throw} may appear without an argument (as shown above), in which case
2588 the object caught by the @code{@@catch} will be rethrown.
2590 Note that only (pointers to) Objective-C objects may be thrown and
2591 caught using this scheme. When an object is thrown, it will be caught
2592 by the nearest @code{@@catch} clause capable of handling objects of that type,
2593 analogously to how @code{catch} blocks work in C++ and Java. A
2594 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
2595 any and all Objective-C exceptions not caught by previous @code{@@catch}
2598 The @code{@@finally} clause, if present, will be executed upon exit from the
2599 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
2600 regardless of whether any exceptions are thrown, caught or rethrown
2601 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
2602 of the @code{finally} clause in Java.
2604 There are several caveats to using the new exception mechanism:
2608 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
2609 idioms provided by the @code{NSException} class, the new
2610 exceptions can only be used on Mac OS X 10.3 (Panther) and later
2611 systems, due to additional functionality needed in the (NeXT) Objective-C
2615 As mentioned above, the new exceptions do not support handling
2616 types other than Objective-C objects. Furthermore, when used from
2617 Objective-C++, the Objective-C exception model does not interoperate with C++
2618 exceptions at this time. This means you cannot @code{@@throw} an exception
2619 from Objective-C and @code{catch} it in C++, or vice versa
2620 (i.e., @code{throw @dots{} @@catch}).
2623 The @option{-fobjc-exceptions} switch also enables the use of synchronization
2624 blocks for thread-safe execution:
2627 @@synchronized (ObjCClass *guard) @{
2632 Upon entering the @code{@@synchronized} block, a thread of execution shall
2633 first check whether a lock has been placed on the corresponding @code{guard}
2634 object by another thread. If it has, the current thread shall wait until
2635 the other thread relinquishes its lock. Once @code{guard} becomes available,
2636 the current thread will place its own lock on it, execute the code contained in
2637 the @code{@@synchronized} block, and finally relinquish the lock (thereby
2638 making @code{guard} available to other threads).
2640 Unlike Java, Objective-C does not allow for entire methods to be marked
2641 @code{@@synchronized}. Note that throwing exceptions out of
2642 @code{@@synchronized} blocks is allowed, and will cause the guarding object
2643 to be unlocked properly.
2647 Enable garbage collection (GC) in Objective-C and Objective-C++ programs.
2649 @item -freplace-objc-classes
2650 @opindex freplace-objc-classes
2651 Emit a special marker instructing @command{ld(1)} not to statically link in
2652 the resulting object file, and allow @command{dyld(1)} to load it in at
2653 run time instead. This is used in conjunction with the Fix-and-Continue
2654 debugging mode, where the object file in question may be recompiled and
2655 dynamically reloaded in the course of program execution, without the need
2656 to restart the program itself. Currently, Fix-and-Continue functionality
2657 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
2662 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
2663 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
2664 compile time) with static class references that get initialized at load time,
2665 which improves run-time performance. Specifying the @option{-fzero-link} flag
2666 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
2667 to be retained. This is useful in Zero-Link debugging mode, since it allows
2668 for individual class implementations to be modified during program execution.
2672 Dump interface declarations for all classes seen in the source file to a
2673 file named @file{@var{sourcename}.decl}.
2675 @item -Wassign-intercept @r{(Objective-C and Objective-C++ only)}
2676 @opindex Wassign-intercept
2677 @opindex Wno-assign-intercept
2678 Warn whenever an Objective-C assignment is being intercepted by the
2681 @item -Wno-protocol @r{(Objective-C and Objective-C++ only)}
2682 @opindex Wno-protocol
2684 If a class is declared to implement a protocol, a warning is issued for
2685 every method in the protocol that is not implemented by the class. The
2686 default behavior is to issue a warning for every method not explicitly
2687 implemented in the class, even if a method implementation is inherited
2688 from the superclass. If you use the @option{-Wno-protocol} option, then
2689 methods inherited from the superclass are considered to be implemented,
2690 and no warning is issued for them.
2692 @item -Wselector @r{(Objective-C and Objective-C++ only)}
2694 @opindex Wno-selector
2695 Warn if multiple methods of different types for the same selector are
2696 found during compilation. The check is performed on the list of methods
2697 in the final stage of compilation. Additionally, a check is performed
2698 for each selector appearing in a @code{@@selector(@dots{})}
2699 expression, and a corresponding method for that selector has been found
2700 during compilation. Because these checks scan the method table only at
2701 the end of compilation, these warnings are not produced if the final
2702 stage of compilation is not reached, for example because an error is
2703 found during compilation, or because the @option{-fsyntax-only} option is
2706 @item -Wstrict-selector-match @r{(Objective-C and Objective-C++ only)}
2707 @opindex Wstrict-selector-match
2708 @opindex Wno-strict-selector-match
2709 Warn if multiple methods with differing argument and/or return types are
2710 found for a given selector when attempting to send a message using this
2711 selector to a receiver of type @code{id} or @code{Class}. When this flag
2712 is off (which is the default behavior), the compiler will omit such warnings
2713 if any differences found are confined to types which share the same size
2716 @item -Wundeclared-selector @r{(Objective-C and Objective-C++ only)}
2717 @opindex Wundeclared-selector
2718 @opindex Wno-undeclared-selector
2719 Warn if a @code{@@selector(@dots{})} expression referring to an
2720 undeclared selector is found. A selector is considered undeclared if no
2721 method with that name has been declared before the
2722 @code{@@selector(@dots{})} expression, either explicitly in an
2723 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
2724 an @code{@@implementation} section. This option always performs its
2725 checks as soon as a @code{@@selector(@dots{})} expression is found,
2726 while @option{-Wselector} only performs its checks in the final stage of
2727 compilation. This also enforces the coding style convention
2728 that methods and selectors must be declared before being used.
2730 @item -print-objc-runtime-info
2731 @opindex print-objc-runtime-info
2732 Generate C header describing the largest structure that is passed by
2737 @node Language Independent Options
2738 @section Options to Control Diagnostic Messages Formatting
2739 @cindex options to control diagnostics formatting
2740 @cindex diagnostic messages
2741 @cindex message formatting
2743 Traditionally, diagnostic messages have been formatted irrespective of
2744 the output device's aspect (e.g.@: its width, @dots{}). The options described
2745 below can be used to control the diagnostic messages formatting
2746 algorithm, e.g.@: how many characters per line, how often source location
2747 information should be reported. Right now, only the C++ front end can
2748 honor these options. However it is expected, in the near future, that
2749 the remaining front ends would be able to digest them correctly.
2752 @item -fmessage-length=@var{n}
2753 @opindex fmessage-length
2754 Try to format error messages so that they fit on lines of about @var{n}
2755 characters. The default is 72 characters for @command{g++} and 0 for the rest of
2756 the front ends supported by GCC@. If @var{n} is zero, then no
2757 line-wrapping will be done; each error message will appear on a single
2760 @opindex fdiagnostics-show-location
2761 @item -fdiagnostics-show-location=once
2762 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
2763 reporter to emit @emph{once} source location information; that is, in
2764 case the message is too long to fit on a single physical line and has to
2765 be wrapped, the source location won't be emitted (as prefix) again,
2766 over and over, in subsequent continuation lines. This is the default
2769 @item -fdiagnostics-show-location=every-line
2770 Only meaningful in line-wrapping mode. Instructs the diagnostic
2771 messages reporter to emit the same source location information (as
2772 prefix) for physical lines that result from the process of breaking
2773 a message which is too long to fit on a single line.
2775 @item -fdiagnostics-show-option
2776 @opindex fdiagnostics-show-option
2777 This option instructs the diagnostic machinery to add text to each
2778 diagnostic emitted, which indicates which command line option directly
2779 controls that diagnostic, when such an option is known to the
2780 diagnostic machinery.
2782 @item -Wcoverage-mismatch
2783 @opindex Wcoverage-mismatch
2784 Warn if feedback profiles do not match when using the
2785 @option{-fprofile-use} option.
2786 If a source file was changed between @option{-fprofile-gen} and
2787 @option{-fprofile-use}, the files with the profile feedback can fail
2788 to match the source file and GCC can not use the profile feedback
2789 information. By default, this warning is enabled and is treated as an
2790 error. @option{-Wno-coverage-mismatch} can be used to disable the
2791 warning or @option{-Wno-error=coverage-mismatch} can be used to
2792 disable the error. Disable the error for this warning can result in
2793 poorly optimized code, so disabling the error is useful only in the
2794 case of very minor changes such as bug fixes to an existing code-base.
2795 Completely disabling the warning is not recommended.
2799 @node Warning Options
2800 @section Options to Request or Suppress Warnings
2801 @cindex options to control warnings
2802 @cindex warning messages
2803 @cindex messages, warning
2804 @cindex suppressing warnings
2806 Warnings are diagnostic messages that report constructions which
2807 are not inherently erroneous but which are risky or suggest there
2808 may have been an error.
2810 The following language-independent options do not enable specific
2811 warnings but control the kinds of diagnostics produced by GCC.
2814 @cindex syntax checking
2816 @opindex fsyntax-only
2817 Check the code for syntax errors, but don't do anything beyond that.
2821 Inhibit all warning messages.
2826 Make all warnings into errors.
2831 Make the specified warning into an error. The specifier for a warning
2832 is appended, for example @option{-Werror=switch} turns the warnings
2833 controlled by @option{-Wswitch} into errors. This switch takes a
2834 negative form, to be used to negate @option{-Werror} for specific
2835 warnings, for example @option{-Wno-error=switch} makes
2836 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
2837 is in effect. You can use the @option{-fdiagnostics-show-option}
2838 option to have each controllable warning amended with the option which
2839 controls it, to determine what to use with this option.
2841 Note that specifying @option{-Werror=}@var{foo} automatically implies
2842 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
2845 @item -Wfatal-errors
2846 @opindex Wfatal-errors
2847 @opindex Wno-fatal-errors
2848 This option causes the compiler to abort compilation on the first error
2849 occurred rather than trying to keep going and printing further error
2854 You can request many specific warnings with options beginning
2855 @samp{-W}, for example @option{-Wimplicit} to request warnings on
2856 implicit declarations. Each of these specific warning options also
2857 has a negative form beginning @samp{-Wno-} to turn off warnings; for
2858 example, @option{-Wno-implicit}. This manual lists only one of the
2859 two forms, whichever is not the default. For further,
2860 language-specific options also refer to @ref{C++ Dialect Options} and
2861 @ref{Objective-C and Objective-C++ Dialect Options}.
2863 When an unrecognized warning label is requested (e.g.,
2864 @option{-Wunknown-warning}), GCC will emit a diagnostic stating
2865 that the option is not recognized. However, if the @samp{-Wno-} form
2866 is used, the behavior is slightly different: No diagnostic will be
2867 produced for @option{-Wno-unknown-warning} unless other diagnostics
2868 are being produced. This allows the use of new @option{-Wno-} options
2869 with old compilers, but if something goes wrong, the compiler will
2870 warn that an unrecognized option was used.
2875 Issue all the warnings demanded by strict ISO C and ISO C++;
2876 reject all programs that use forbidden extensions, and some other
2877 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2878 version of the ISO C standard specified by any @option{-std} option used.
2880 Valid ISO C and ISO C++ programs should compile properly with or without
2881 this option (though a rare few will require @option{-ansi} or a
2882 @option{-std} option specifying the required version of ISO C)@. However,
2883 without this option, certain GNU extensions and traditional C and C++
2884 features are supported as well. With this option, they are rejected.
2886 @option{-pedantic} does not cause warning messages for use of the
2887 alternate keywords whose names begin and end with @samp{__}. Pedantic
2888 warnings are also disabled in the expression that follows
2889 @code{__extension__}. However, only system header files should use
2890 these escape routes; application programs should avoid them.
2891 @xref{Alternate Keywords}.
2893 Some users try to use @option{-pedantic} to check programs for strict ISO
2894 C conformance. They soon find that it does not do quite what they want:
2895 it finds some non-ISO practices, but not all---only those for which
2896 ISO C @emph{requires} a diagnostic, and some others for which
2897 diagnostics have been added.
2899 A feature to report any failure to conform to ISO C might be useful in
2900 some instances, but would require considerable additional work and would
2901 be quite different from @option{-pedantic}. We don't have plans to
2902 support such a feature in the near future.
2904 Where the standard specified with @option{-std} represents a GNU
2905 extended dialect of C, such as @samp{gnu90} or @samp{gnu99}, there is a
2906 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2907 extended dialect is based. Warnings from @option{-pedantic} are given
2908 where they are required by the base standard. (It would not make sense
2909 for such warnings to be given only for features not in the specified GNU
2910 C dialect, since by definition the GNU dialects of C include all
2911 features the compiler supports with the given option, and there would be
2912 nothing to warn about.)
2914 @item -pedantic-errors
2915 @opindex pedantic-errors
2916 Like @option{-pedantic}, except that errors are produced rather than
2922 This enables all the warnings about constructions that some users
2923 consider questionable, and that are easy to avoid (or modify to
2924 prevent the warning), even in conjunction with macros. This also
2925 enables some language-specific warnings described in @ref{C++ Dialect
2926 Options} and @ref{Objective-C and Objective-C++ Dialect Options}.
2928 @option{-Wall} turns on the following warning flags:
2930 @gccoptlist{-Waddress @gol
2931 -Warray-bounds @r{(only with} @option{-O2}@r{)} @gol
2933 -Wchar-subscripts @gol
2934 -Wenum-compare @r{(in C/Objc; this is on by default in C++)} @gol
2935 -Wimplicit-int @r{(C and Objective-C only)} @gol
2936 -Wimplicit-function-declaration @r{(C and Objective-C only)} @gol
2939 -Wmain @r{(only for C/ObjC and unless} @option{-ffreestanding}@r{)} @gol
2940 -Wmissing-braces @gol
2946 -Wsequence-point @gol
2947 -Wsign-compare @r{(only in C++)} @gol
2948 -Wstrict-aliasing @gol
2949 -Wstrict-overflow=1 @gol
2952 -Wuninitialized @gol
2953 -Wunknown-pragmas @gol
2954 -Wunused-function @gol
2957 -Wunused-variable @gol
2958 -Wvolatile-register-var @gol
2961 Note that some warning flags are not implied by @option{-Wall}. Some of
2962 them warn about constructions that users generally do not consider
2963 questionable, but which occasionally you might wish to check for;
2964 others warn about constructions that are necessary or hard to avoid in
2965 some cases, and there is no simple way to modify the code to suppress
2966 the warning. Some of them are enabled by @option{-Wextra} but many of
2967 them must be enabled individually.
2973 This enables some extra warning flags that are not enabled by
2974 @option{-Wall}. (This option used to be called @option{-W}. The older
2975 name is still supported, but the newer name is more descriptive.)
2977 @gccoptlist{-Wclobbered @gol
2979 -Wignored-qualifiers @gol
2980 -Wmissing-field-initializers @gol
2981 -Wmissing-parameter-type @r{(C only)} @gol
2982 -Wold-style-declaration @r{(C only)} @gol
2983 -Woverride-init @gol
2986 -Wuninitialized @gol
2987 -Wunused-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)} @gol
2988 -Wunused-but-set-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)} @gol
2991 The option @option{-Wextra} also prints warning messages for the
2997 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2998 @samp{>}, or @samp{>=}.
3001 (C++ only) An enumerator and a non-enumerator both appear in a
3002 conditional expression.
3005 (C++ only) Ambiguous virtual bases.
3008 (C++ only) Subscripting an array which has been declared @samp{register}.
3011 (C++ only) Taking the address of a variable which has been declared
3015 (C++ only) A base class is not initialized in a derived class' copy
3020 @item -Wchar-subscripts
3021 @opindex Wchar-subscripts
3022 @opindex Wno-char-subscripts
3023 Warn if an array subscript has type @code{char}. This is a common cause
3024 of error, as programmers often forget that this type is signed on some
3026 This warning is enabled by @option{-Wall}.
3030 @opindex Wno-comment
3031 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
3032 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
3033 This warning is enabled by @option{-Wall}.
3036 @r{(C, Objective-C, C++, Objective-C++ and Fortran only)}
3038 Suppress warning messages emitted by @code{#warning} directives.
3043 @opindex ffreestanding
3044 @opindex fno-builtin
3045 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
3046 the arguments supplied have types appropriate to the format string
3047 specified, and that the conversions specified in the format string make
3048 sense. This includes standard functions, and others specified by format
3049 attributes (@pxref{Function Attributes}), in the @code{printf},
3050 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
3051 not in the C standard) families (or other target-specific families).
3052 Which functions are checked without format attributes having been
3053 specified depends on the standard version selected, and such checks of
3054 functions without the attribute specified are disabled by
3055 @option{-ffreestanding} or @option{-fno-builtin}.
3057 The formats are checked against the format features supported by GNU
3058 libc version 2.2. These include all ISO C90 and C99 features, as well
3059 as features from the Single Unix Specification and some BSD and GNU
3060 extensions. Other library implementations may not support all these
3061 features; GCC does not support warning about features that go beyond a
3062 particular library's limitations. However, if @option{-pedantic} is used
3063 with @option{-Wformat}, warnings will be given about format features not
3064 in the selected standard version (but not for @code{strfmon} formats,
3065 since those are not in any version of the C standard). @xref{C Dialect
3066 Options,,Options Controlling C Dialect}.
3068 Since @option{-Wformat} also checks for null format arguments for
3069 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
3071 @option{-Wformat} is included in @option{-Wall}. For more control over some
3072 aspects of format checking, the options @option{-Wformat-y2k},
3073 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
3074 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
3075 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
3078 @opindex Wformat-y2k
3079 @opindex Wno-format-y2k
3080 If @option{-Wformat} is specified, also warn about @code{strftime}
3081 formats which may yield only a two-digit year.
3083 @item -Wno-format-contains-nul
3084 @opindex Wno-format-contains-nul
3085 @opindex Wformat-contains-nul
3086 If @option{-Wformat} is specified, do not warn about format strings that
3089 @item -Wno-format-extra-args
3090 @opindex Wno-format-extra-args
3091 @opindex Wformat-extra-args
3092 If @option{-Wformat} is specified, do not warn about excess arguments to a
3093 @code{printf} or @code{scanf} format function. The C standard specifies
3094 that such arguments are ignored.
3096 Where the unused arguments lie between used arguments that are
3097 specified with @samp{$} operand number specifications, normally
3098 warnings are still given, since the implementation could not know what
3099 type to pass to @code{va_arg} to skip the unused arguments. However,
3100 in the case of @code{scanf} formats, this option will suppress the
3101 warning if the unused arguments are all pointers, since the Single
3102 Unix Specification says that such unused arguments are allowed.
3104 @item -Wno-format-zero-length @r{(C and Objective-C only)}
3105 @opindex Wno-format-zero-length
3106 @opindex Wformat-zero-length
3107 If @option{-Wformat} is specified, do not warn about zero-length formats.
3108 The C standard specifies that zero-length formats are allowed.
3110 @item -Wformat-nonliteral
3111 @opindex Wformat-nonliteral
3112 @opindex Wno-format-nonliteral
3113 If @option{-Wformat} is specified, also warn if the format string is not a
3114 string literal and so cannot be checked, unless the format function
3115 takes its format arguments as a @code{va_list}.
3117 @item -Wformat-security
3118 @opindex Wformat-security
3119 @opindex Wno-format-security
3120 If @option{-Wformat} is specified, also warn about uses of format
3121 functions that represent possible security problems. At present, this
3122 warns about calls to @code{printf} and @code{scanf} functions where the
3123 format string is not a string literal and there are no format arguments,
3124 as in @code{printf (foo);}. This may be a security hole if the format
3125 string came from untrusted input and contains @samp{%n}. (This is
3126 currently a subset of what @option{-Wformat-nonliteral} warns about, but
3127 in future warnings may be added to @option{-Wformat-security} that are not
3128 included in @option{-Wformat-nonliteral}.)
3132 @opindex Wno-format=2
3133 Enable @option{-Wformat} plus format checks not included in
3134 @option{-Wformat}. Currently equivalent to @samp{-Wformat
3135 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
3137 @item -Wnonnull @r{(C and Objective-C only)}
3139 @opindex Wno-nonnull
3140 Warn about passing a null pointer for arguments marked as
3141 requiring a non-null value by the @code{nonnull} function attribute.
3143 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
3144 can be disabled with the @option{-Wno-nonnull} option.
3146 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
3148 @opindex Wno-init-self
3149 Warn about uninitialized variables which are initialized with themselves.
3150 Note this option can only be used with the @option{-Wuninitialized} option.
3152 For example, GCC will warn about @code{i} being uninitialized in the
3153 following snippet only when @option{-Winit-self} has been specified:
3164 @item -Wimplicit-int @r{(C and Objective-C only)}
3165 @opindex Wimplicit-int
3166 @opindex Wno-implicit-int
3167 Warn when a declaration does not specify a type.
3168 This warning is enabled by @option{-Wall}.
3170 @item -Wimplicit-function-declaration @r{(C and Objective-C only)}
3171 @opindex Wimplicit-function-declaration
3172 @opindex Wno-implicit-function-declaration
3173 Give a warning whenever a function is used before being declared. In
3174 C99 mode (@option{-std=c99} or @option{-std=gnu99}), this warning is
3175 enabled by default and it is made into an error by
3176 @option{-pedantic-errors}. This warning is also enabled by
3179 @item -Wimplicit @r{(C and Objective-C only)}
3181 @opindex Wno-implicit
3182 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
3183 This warning is enabled by @option{-Wall}.
3185 @item -Wignored-qualifiers @r{(C and C++ only)}
3186 @opindex Wignored-qualifiers
3187 @opindex Wno-ignored-qualifiers
3188 Warn if the return type of a function has a type qualifier
3189 such as @code{const}. For ISO C such a type qualifier has no effect,
3190 since the value returned by a function is not an lvalue.
3191 For C++, the warning is only emitted for scalar types or @code{void}.
3192 ISO C prohibits qualified @code{void} return types on function
3193 definitions, so such return types always receive a warning
3194 even without this option.
3196 This warning is also enabled by @option{-Wextra}.
3201 Warn if the type of @samp{main} is suspicious. @samp{main} should be
3202 a function with external linkage, returning int, taking either zero
3203 arguments, two, or three arguments of appropriate types. This warning
3204 is enabled by default in C++ and is enabled by either @option{-Wall}
3205 or @option{-pedantic}.
3207 @item -Wmissing-braces
3208 @opindex Wmissing-braces
3209 @opindex Wno-missing-braces
3210 Warn if an aggregate or union initializer is not fully bracketed. In
3211 the following example, the initializer for @samp{a} is not fully
3212 bracketed, but that for @samp{b} is fully bracketed.
3215 int a[2][2] = @{ 0, 1, 2, 3 @};
3216 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
3219 This warning is enabled by @option{-Wall}.
3221 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
3222 @opindex Wmissing-include-dirs
3223 @opindex Wno-missing-include-dirs
3224 Warn if a user-supplied include directory does not exist.
3227 @opindex Wparentheses
3228 @opindex Wno-parentheses
3229 Warn if parentheses are omitted in certain contexts, such
3230 as when there is an assignment in a context where a truth value
3231 is expected, or when operators are nested whose precedence people
3232 often get confused about.
3234 Also warn if a comparison like @samp{x<=y<=z} appears; this is
3235 equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different
3236 interpretation from that of ordinary mathematical notation.
3238 Also warn about constructions where there may be confusion to which
3239 @code{if} statement an @code{else} branch belongs. Here is an example of
3254 In C/C++, every @code{else} branch belongs to the innermost possible
3255 @code{if} statement, which in this example is @code{if (b)}. This is
3256 often not what the programmer expected, as illustrated in the above
3257 example by indentation the programmer chose. When there is the
3258 potential for this confusion, GCC will issue a warning when this flag
3259 is specified. To eliminate the warning, add explicit braces around
3260 the innermost @code{if} statement so there is no way the @code{else}
3261 could belong to the enclosing @code{if}. The resulting code would
3278 Also warn for dangerous uses of the
3279 ?: with omitted middle operand GNU extension. When the condition
3280 in the ?: operator is a boolean expression the omitted value will
3281 be always 1. Often the user expects it to be a value computed
3282 inside the conditional expression instead.
3284 This warning is enabled by @option{-Wall}.
3286 @item -Wsequence-point
3287 @opindex Wsequence-point
3288 @opindex Wno-sequence-point
3289 Warn about code that may have undefined semantics because of violations
3290 of sequence point rules in the C and C++ standards.
3292 The C and C++ standards defines the order in which expressions in a C/C++
3293 program are evaluated in terms of @dfn{sequence points}, which represent
3294 a partial ordering between the execution of parts of the program: those
3295 executed before the sequence point, and those executed after it. These
3296 occur after the evaluation of a full expression (one which is not part
3297 of a larger expression), after the evaluation of the first operand of a
3298 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
3299 function is called (but after the evaluation of its arguments and the
3300 expression denoting the called function), and in certain other places.
3301 Other than as expressed by the sequence point rules, the order of
3302 evaluation of subexpressions of an expression is not specified. All
3303 these rules describe only a partial order rather than a total order,
3304 since, for example, if two functions are called within one expression
3305 with no sequence point between them, the order in which the functions
3306 are called is not specified. However, the standards committee have
3307 ruled that function calls do not overlap.
3309 It is not specified when between sequence points modifications to the
3310 values of objects take effect. Programs whose behavior depends on this
3311 have undefined behavior; the C and C++ standards specify that ``Between
3312 the previous and next sequence point an object shall have its stored
3313 value modified at most once by the evaluation of an expression.
3314 Furthermore, the prior value shall be read only to determine the value
3315 to be stored.''. If a program breaks these rules, the results on any
3316 particular implementation are entirely unpredictable.
3318 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
3319 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
3320 diagnosed by this option, and it may give an occasional false positive
3321 result, but in general it has been found fairly effective at detecting
3322 this sort of problem in programs.
3324 The standard is worded confusingly, therefore there is some debate
3325 over the precise meaning of the sequence point rules in subtle cases.
3326 Links to discussions of the problem, including proposed formal
3327 definitions, may be found on the GCC readings page, at
3328 @w{@uref{http://gcc.gnu.org/readings.html}}.
3330 This warning is enabled by @option{-Wall} for C and C++.
3333 @opindex Wreturn-type
3334 @opindex Wno-return-type
3335 Warn whenever a function is defined with a return-type that defaults
3336 to @code{int}. Also warn about any @code{return} statement with no
3337 return-value in a function whose return-type is not @code{void}
3338 (falling off the end of the function body is considered returning
3339 without a value), and about a @code{return} statement with an
3340 expression in a function whose return-type is @code{void}.
3342 For C++, a function without return type always produces a diagnostic
3343 message, even when @option{-Wno-return-type} is specified. The only
3344 exceptions are @samp{main} and functions defined in system headers.
3346 This warning is enabled by @option{-Wall}.
3351 Warn whenever a @code{switch} statement has an index of enumerated type
3352 and lacks a @code{case} for one or more of the named codes of that
3353 enumeration. (The presence of a @code{default} label prevents this
3354 warning.) @code{case} labels outside the enumeration range also
3355 provoke warnings when this option is used (even if there is a
3356 @code{default} label).
3357 This warning is enabled by @option{-Wall}.
3359 @item -Wswitch-default
3360 @opindex Wswitch-default
3361 @opindex Wno-switch-default
3362 Warn whenever a @code{switch} statement does not have a @code{default}
3366 @opindex Wswitch-enum
3367 @opindex Wno-switch-enum
3368 Warn whenever a @code{switch} statement has an index of enumerated type
3369 and lacks a @code{case} for one or more of the named codes of that
3370 enumeration. @code{case} labels outside the enumeration range also
3371 provoke warnings when this option is used. The only difference
3372 between @option{-Wswitch} and this option is that this option gives a
3373 warning about an omitted enumeration code even if there is a
3374 @code{default} label.
3376 @item -Wsync-nand @r{(C and C++ only)}
3378 @opindex Wno-sync-nand
3379 Warn when @code{__sync_fetch_and_nand} and @code{__sync_nand_and_fetch}
3380 built-in functions are used. These functions changed semantics in GCC 4.4.
3384 @opindex Wno-trigraphs
3385 Warn if any trigraphs are encountered that might change the meaning of
3386 the program (trigraphs within comments are not warned about).
3387 This warning is enabled by @option{-Wall}.
3389 @item -Wunused-but-set-parameter
3390 @opindex Wunused-but-set-parameter
3391 @opindex Wno-unused-but-set-parameter
3392 Warn whenever a function parameter is assigned to, but otherwise unused
3393 (aside from its declaration).
3395 To suppress this warning use the @samp{unused} attribute
3396 (@pxref{Variable Attributes}).
3398 This warning is also enabled by @option{-Wunused} together with
3401 @item -Wunused-but-set-variable
3402 @opindex Wunused-but-set-variable
3403 @opindex Wno-unused-but-set-variable
3404 Warn whenever a local variable is assigned to, but otherwise unused
3405 (aside from its declaration).
3406 This warning is enabled by @option{-Wall}.
3408 To suppress this warning use the @samp{unused} attribute
3409 (@pxref{Variable Attributes}).
3411 This warning is also enabled by @option{-Wunused}, which is enabled
3414 @item -Wunused-function
3415 @opindex Wunused-function
3416 @opindex Wno-unused-function
3417 Warn whenever a static function is declared but not defined or a
3418 non-inline static function is unused.
3419 This warning is enabled by @option{-Wall}.
3421 @item -Wunused-label
3422 @opindex Wunused-label
3423 @opindex Wno-unused-label
3424 Warn whenever a label is declared but not used.
3425 This warning is enabled by @option{-Wall}.
3427 To suppress this warning use the @samp{unused} attribute
3428 (@pxref{Variable Attributes}).
3430 @item -Wunused-parameter
3431 @opindex Wunused-parameter
3432 @opindex Wno-unused-parameter
3433 Warn whenever a function parameter is unused aside from its declaration.
3435 To suppress this warning use the @samp{unused} attribute
3436 (@pxref{Variable Attributes}).
3438 @item -Wno-unused-result
3439 @opindex Wunused-result
3440 @opindex Wno-unused-result
3441 Do not warn if a caller of a function marked with attribute
3442 @code{warn_unused_result} (@pxref{Variable Attributes}) does not use
3443 its return value. The default is @option{-Wunused-result}.
3445 @item -Wunused-variable
3446 @opindex Wunused-variable
3447 @opindex Wno-unused-variable
3448 Warn whenever a local variable or non-constant static variable is unused
3449 aside from its declaration.
3450 This warning is enabled by @option{-Wall}.
3452 To suppress this warning use the @samp{unused} attribute
3453 (@pxref{Variable Attributes}).
3455 @item -Wunused-value
3456 @opindex Wunused-value
3457 @opindex Wno-unused-value
3458 Warn whenever a statement computes a result that is explicitly not
3459 used. To suppress this warning cast the unused expression to
3460 @samp{void}. This includes an expression-statement or the left-hand
3461 side of a comma expression that contains no side effects. For example,
3462 an expression such as @samp{x[i,j]} will cause a warning, while
3463 @samp{x[(void)i,j]} will not.
3465 This warning is enabled by @option{-Wall}.
3470 All the above @option{-Wunused} options combined.
3472 In order to get a warning about an unused function parameter, you must
3473 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
3474 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
3476 @item -Wuninitialized
3477 @opindex Wuninitialized
3478 @opindex Wno-uninitialized
3479 Warn if an automatic variable is used without first being initialized
3480 or if a variable may be clobbered by a @code{setjmp} call. In C++,
3481 warn if a non-static reference or non-static @samp{const} member
3482 appears in a class without constructors.
3484 If you want to warn about code which uses the uninitialized value of the
3485 variable in its own initializer, use the @option{-Winit-self} option.
3487 These warnings occur for individual uninitialized or clobbered
3488 elements of structure, union or array variables as well as for
3489 variables which are uninitialized or clobbered as a whole. They do
3490 not occur for variables or elements declared @code{volatile}. Because
3491 these warnings depend on optimization, the exact variables or elements
3492 for which there are warnings will depend on the precise optimization
3493 options and version of GCC used.
3495 Note that there may be no warning about a variable that is used only
3496 to compute a value that itself is never used, because such
3497 computations may be deleted by data flow analysis before the warnings
3500 These warnings are made optional because GCC is not smart
3501 enough to see all the reasons why the code might be correct
3502 despite appearing to have an error. Here is one example of how
3523 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
3524 always initialized, but GCC doesn't know this. Here is
3525 another common case:
3530 if (change_y) save_y = y, y = new_y;
3532 if (change_y) y = save_y;
3537 This has no bug because @code{save_y} is used only if it is set.
3539 @cindex @code{longjmp} warnings
3540 This option also warns when a non-volatile automatic variable might be
3541 changed by a call to @code{longjmp}. These warnings as well are possible
3542 only in optimizing compilation.
3544 The compiler sees only the calls to @code{setjmp}. It cannot know
3545 where @code{longjmp} will be called; in fact, a signal handler could
3546 call it at any point in the code. As a result, you may get a warning
3547 even when there is in fact no problem because @code{longjmp} cannot
3548 in fact be called at the place which would cause a problem.
3550 Some spurious warnings can be avoided if you declare all the functions
3551 you use that never return as @code{noreturn}. @xref{Function
3554 This warning is enabled by @option{-Wall} or @option{-Wextra}.
3556 @item -Wunknown-pragmas
3557 @opindex Wunknown-pragmas
3558 @opindex Wno-unknown-pragmas
3559 @cindex warning for unknown pragmas
3560 @cindex unknown pragmas, warning
3561 @cindex pragmas, warning of unknown
3562 Warn when a #pragma directive is encountered which is not understood by
3563 GCC@. If this command line option is used, warnings will even be issued
3564 for unknown pragmas in system header files. This is not the case if
3565 the warnings were only enabled by the @option{-Wall} command line option.
3568 @opindex Wno-pragmas
3570 Do not warn about misuses of pragmas, such as incorrect parameters,
3571 invalid syntax, or conflicts between pragmas. See also
3572 @samp{-Wunknown-pragmas}.
3574 @item -Wstrict-aliasing
3575 @opindex Wstrict-aliasing
3576 @opindex Wno-strict-aliasing
3577 This option is only active when @option{-fstrict-aliasing} is active.
3578 It warns about code which might break the strict aliasing rules that the
3579 compiler is using for optimization. The warning does not catch all
3580 cases, but does attempt to catch the more common pitfalls. It is
3581 included in @option{-Wall}.
3582 It is equivalent to @option{-Wstrict-aliasing=3}
3584 @item -Wstrict-aliasing=n
3585 @opindex Wstrict-aliasing=n
3586 @opindex Wno-strict-aliasing=n
3587 This option is only active when @option{-fstrict-aliasing} is active.
3588 It warns about code which might break the strict aliasing rules that the
3589 compiler is using for optimization.
3590 Higher levels correspond to higher accuracy (fewer false positives).
3591 Higher levels also correspond to more effort, similar to the way -O works.
3592 @option{-Wstrict-aliasing} is equivalent to @option{-Wstrict-aliasing=n},
3595 Level 1: Most aggressive, quick, least accurate.
3596 Possibly useful when higher levels
3597 do not warn but -fstrict-aliasing still breaks the code, as it has very few
3598 false negatives. However, it has many false positives.
3599 Warns for all pointer conversions between possibly incompatible types,
3600 even if never dereferenced. Runs in the frontend only.
3602 Level 2: Aggressive, quick, not too precise.
3603 May still have many false positives (not as many as level 1 though),
3604 and few false negatives (but possibly more than level 1).
3605 Unlike level 1, it only warns when an address is taken. Warns about
3606 incomplete types. Runs in the frontend only.
3608 Level 3 (default for @option{-Wstrict-aliasing}):
3609 Should have very few false positives and few false
3610 negatives. Slightly slower than levels 1 or 2 when optimization is enabled.
3611 Takes care of the common pun+dereference pattern in the frontend:
3612 @code{*(int*)&some_float}.
3613 If optimization is enabled, it also runs in the backend, where it deals
3614 with multiple statement cases using flow-sensitive points-to information.
3615 Only warns when the converted pointer is dereferenced.
3616 Does not warn about incomplete types.
3618 @item -Wstrict-overflow
3619 @itemx -Wstrict-overflow=@var{n}
3620 @opindex Wstrict-overflow
3621 @opindex Wno-strict-overflow
3622 This option is only active when @option{-fstrict-overflow} is active.
3623 It warns about cases where the compiler optimizes based on the
3624 assumption that signed overflow does not occur. Note that it does not
3625 warn about all cases where the code might overflow: it only warns
3626 about cases where the compiler implements some optimization. Thus
3627 this warning depends on the optimization level.
3629 An optimization which assumes that signed overflow does not occur is
3630 perfectly safe if the values of the variables involved are such that
3631 overflow never does, in fact, occur. Therefore this warning can
3632 easily give a false positive: a warning about code which is not
3633 actually a problem. To help focus on important issues, several
3634 warning levels are defined. No warnings are issued for the use of
3635 undefined signed overflow when estimating how many iterations a loop
3636 will require, in particular when determining whether a loop will be
3640 @item -Wstrict-overflow=1
3641 Warn about cases which are both questionable and easy to avoid. For
3642 example: @code{x + 1 > x}; with @option{-fstrict-overflow}, the
3643 compiler will simplify this to @code{1}. This level of
3644 @option{-Wstrict-overflow} is enabled by @option{-Wall}; higher levels
3645 are not, and must be explicitly requested.
3647 @item -Wstrict-overflow=2
3648 Also warn about other cases where a comparison is simplified to a
3649 constant. For example: @code{abs (x) >= 0}. This can only be
3650 simplified when @option{-fstrict-overflow} is in effect, because
3651 @code{abs (INT_MIN)} overflows to @code{INT_MIN}, which is less than
3652 zero. @option{-Wstrict-overflow} (with no level) is the same as
3653 @option{-Wstrict-overflow=2}.
3655 @item -Wstrict-overflow=3
3656 Also warn about other cases where a comparison is simplified. For
3657 example: @code{x + 1 > 1} will be simplified to @code{x > 0}.
3659 @item -Wstrict-overflow=4
3660 Also warn about other simplifications not covered by the above cases.
3661 For example: @code{(x * 10) / 5} will be simplified to @code{x * 2}.
3663 @item -Wstrict-overflow=5
3664 Also warn about cases where the compiler reduces the magnitude of a
3665 constant involved in a comparison. For example: @code{x + 2 > y} will
3666 be simplified to @code{x + 1 >= y}. This is reported only at the
3667 highest warning level because this simplification applies to many
3668 comparisons, so this warning level will give a very large number of
3672 @item -Wsuggest-attribute=@r{[}const@r{|}pure@r{|}noreturn@r{]}
3673 @opindex Wsuggest-attribute=
3674 @opindex Wno-suggest-attribute=
3675 Warn for cases where adding an attribute may be beneficial. The
3676 attributes currently supported are listed below.
3679 @item -Wsuggest-attribute=pure
3680 @itemx -Wsuggest-attribute=const
3681 @opindex Wsuggest-attribute=pure
3682 @opindex Wno-suggest-attribute=pure
3683 @opindex Wsuggest-attribute=const
3684 @opindex Wno-suggest-attribute=const
3685 @opindex Wsuggest-attribute=noreturn
3686 @opindex Wno-suggest-attribute=noreturn
3688 Warn about functions which might be candidates for attributes
3689 @code{pure}, @code{const} or @code{noreturn}. The compiler only warns for
3690 functions visible in other compilation units or (in the case of @code{pure} and
3691 @code{const}) if it cannot prove that the function returns normally. A function
3692 returns normally if it doesn't contain an infinite loop nor returns abnormally
3693 by throwing, calling @code{abort()} or trapping. This analysis requires option
3694 @option{-fipa-pure-const}, which is enabled by default at @option{-O} and
3695 higher. Higher optimization levels improve the accuracy of the analysis.
3698 @item -Warray-bounds
3699 @opindex Wno-array-bounds
3700 @opindex Warray-bounds
3701 This option is only active when @option{-ftree-vrp} is active
3702 (default for @option{-O2} and above). It warns about subscripts to arrays
3703 that are always out of bounds. This warning is enabled by @option{-Wall}.
3705 @item -Wno-div-by-zero
3706 @opindex Wno-div-by-zero
3707 @opindex Wdiv-by-zero
3708 Do not warn about compile-time integer division by zero. Floating point
3709 division by zero is not warned about, as it can be a legitimate way of
3710 obtaining infinities and NaNs.
3712 @item -Wsystem-headers
3713 @opindex Wsystem-headers
3714 @opindex Wno-system-headers
3715 @cindex warnings from system headers
3716 @cindex system headers, warnings from
3717 Print warning messages for constructs found in system header files.
3718 Warnings from system headers are normally suppressed, on the assumption
3719 that they usually do not indicate real problems and would only make the
3720 compiler output harder to read. Using this command line option tells
3721 GCC to emit warnings from system headers as if they occurred in user
3722 code. However, note that using @option{-Wall} in conjunction with this
3723 option will @emph{not} warn about unknown pragmas in system
3724 headers---for that, @option{-Wunknown-pragmas} must also be used.
3727 @opindex Wfloat-equal
3728 @opindex Wno-float-equal
3729 Warn if floating point values are used in equality comparisons.
3731 The idea behind this is that sometimes it is convenient (for the
3732 programmer) to consider floating-point values as approximations to
3733 infinitely precise real numbers. If you are doing this, then you need
3734 to compute (by analyzing the code, or in some other way) the maximum or
3735 likely maximum error that the computation introduces, and allow for it
3736 when performing comparisons (and when producing output, but that's a
3737 different problem). In particular, instead of testing for equality, you
3738 would check to see whether the two values have ranges that overlap; and
3739 this is done with the relational operators, so equality comparisons are
3742 @item -Wtraditional @r{(C and Objective-C only)}
3743 @opindex Wtraditional
3744 @opindex Wno-traditional
3745 Warn about certain constructs that behave differently in traditional and
3746 ISO C@. Also warn about ISO C constructs that have no traditional C
3747 equivalent, and/or problematic constructs which should be avoided.
3751 Macro parameters that appear within string literals in the macro body.
3752 In traditional C macro replacement takes place within string literals,
3753 but does not in ISO C@.
3756 In traditional C, some preprocessor directives did not exist.
3757 Traditional preprocessors would only consider a line to be a directive
3758 if the @samp{#} appeared in column 1 on the line. Therefore
3759 @option{-Wtraditional} warns about directives that traditional C
3760 understands but would ignore because the @samp{#} does not appear as the
3761 first character on the line. It also suggests you hide directives like
3762 @samp{#pragma} not understood by traditional C by indenting them. Some
3763 traditional implementations would not recognize @samp{#elif}, so it
3764 suggests avoiding it altogether.
3767 A function-like macro that appears without arguments.
3770 The unary plus operator.
3773 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
3774 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
3775 constants.) Note, these suffixes appear in macros defined in the system
3776 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
3777 Use of these macros in user code might normally lead to spurious
3778 warnings, however GCC's integrated preprocessor has enough context to
3779 avoid warning in these cases.
3782 A function declared external in one block and then used after the end of
3786 A @code{switch} statement has an operand of type @code{long}.
3789 A non-@code{static} function declaration follows a @code{static} one.
3790 This construct is not accepted by some traditional C compilers.
3793 The ISO type of an integer constant has a different width or
3794 signedness from its traditional type. This warning is only issued if
3795 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
3796 typically represent bit patterns, are not warned about.
3799 Usage of ISO string concatenation is detected.
3802 Initialization of automatic aggregates.
3805 Identifier conflicts with labels. Traditional C lacks a separate
3806 namespace for labels.
3809 Initialization of unions. If the initializer is zero, the warning is
3810 omitted. This is done under the assumption that the zero initializer in
3811 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
3812 initializer warnings and relies on default initialization to zero in the
3816 Conversions by prototypes between fixed/floating point values and vice
3817 versa. The absence of these prototypes when compiling with traditional
3818 C would cause serious problems. This is a subset of the possible
3819 conversion warnings, for the full set use @option{-Wtraditional-conversion}.
3822 Use of ISO C style function definitions. This warning intentionally is
3823 @emph{not} issued for prototype declarations or variadic functions
3824 because these ISO C features will appear in your code when using
3825 libiberty's traditional C compatibility macros, @code{PARAMS} and
3826 @code{VPARAMS}. This warning is also bypassed for nested functions
3827 because that feature is already a GCC extension and thus not relevant to
3828 traditional C compatibility.
3831 @item -Wtraditional-conversion @r{(C and Objective-C only)}
3832 @opindex Wtraditional-conversion
3833 @opindex Wno-traditional-conversion
3834 Warn if a prototype causes a type conversion that is different from what
3835 would happen to the same argument in the absence of a prototype. This
3836 includes conversions of fixed point to floating and vice versa, and
3837 conversions changing the width or signedness of a fixed point argument
3838 except when the same as the default promotion.
3840 @item -Wdeclaration-after-statement @r{(C and Objective-C only)}
3841 @opindex Wdeclaration-after-statement
3842 @opindex Wno-declaration-after-statement
3843 Warn when a declaration is found after a statement in a block. This
3844 construct, known from C++, was introduced with ISO C99 and is by default
3845 allowed in GCC@. It is not supported by ISO C90 and was not supported by
3846 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
3851 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
3853 @item -Wno-endif-labels
3854 @opindex Wno-endif-labels
3855 @opindex Wendif-labels
3856 Do not warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
3861 Warn whenever a local variable shadows another local variable, parameter or
3862 global variable or whenever a built-in function is shadowed.
3864 @item -Wlarger-than=@var{len}
3865 @opindex Wlarger-than=@var{len}
3866 @opindex Wlarger-than-@var{len}
3867 Warn whenever an object of larger than @var{len} bytes is defined.
3869 @item -Wframe-larger-than=@var{len}
3870 @opindex Wframe-larger-than
3871 Warn if the size of a function frame is larger than @var{len} bytes.
3872 The computation done to determine the stack frame size is approximate
3873 and not conservative.
3874 The actual requirements may be somewhat greater than @var{len}
3875 even if you do not get a warning. In addition, any space allocated
3876 via @code{alloca}, variable-length arrays, or related constructs
3877 is not included by the compiler when determining
3878 whether or not to issue a warning.
3880 @item -Wunsafe-loop-optimizations
3881 @opindex Wunsafe-loop-optimizations
3882 @opindex Wno-unsafe-loop-optimizations
3883 Warn if the loop cannot be optimized because the compiler could not
3884 assume anything on the bounds of the loop indices. With
3885 @option{-funsafe-loop-optimizations} warn if the compiler made
3888 @item -Wno-pedantic-ms-format @r{(MinGW targets only)}
3889 @opindex Wno-pedantic-ms-format
3890 @opindex Wpedantic-ms-format
3891 Disables the warnings about non-ISO @code{printf} / @code{scanf} format
3892 width specifiers @code{I32}, @code{I64}, and @code{I} used on Windows targets
3893 depending on the MS runtime, when you are using the options @option{-Wformat}
3894 and @option{-pedantic} without gnu-extensions.
3896 @item -Wpointer-arith
3897 @opindex Wpointer-arith
3898 @opindex Wno-pointer-arith
3899 Warn about anything that depends on the ``size of'' a function type or
3900 of @code{void}. GNU C assigns these types a size of 1, for
3901 convenience in calculations with @code{void *} pointers and pointers
3902 to functions. In C++, warn also when an arithmetic operation involves
3903 @code{NULL}. This warning is also enabled by @option{-pedantic}.
3906 @opindex Wtype-limits
3907 @opindex Wno-type-limits
3908 Warn if a comparison is always true or always false due to the limited
3909 range of the data type, but do not warn for constant expressions. For
3910 example, warn if an unsigned variable is compared against zero with
3911 @samp{<} or @samp{>=}. This warning is also enabled by
3914 @item -Wbad-function-cast @r{(C and Objective-C only)}
3915 @opindex Wbad-function-cast
3916 @opindex Wno-bad-function-cast
3917 Warn whenever a function call is cast to a non-matching type.
3918 For example, warn if @code{int malloc()} is cast to @code{anything *}.
3920 @item -Wc++-compat @r{(C and Objective-C only)}
3921 Warn about ISO C constructs that are outside of the common subset of
3922 ISO C and ISO C++, e.g.@: request for implicit conversion from
3923 @code{void *} to a pointer to non-@code{void} type.
3925 @item -Wc++0x-compat @r{(C++ and Objective-C++ only)}
3926 Warn about C++ constructs whose meaning differs between ISO C++ 1998 and
3927 ISO C++ 200x, e.g., identifiers in ISO C++ 1998 that will become keywords
3928 in ISO C++ 200x. This warning is enabled by @option{-Wall}.
3932 @opindex Wno-cast-qual
3933 Warn whenever a pointer is cast so as to remove a type qualifier from
3934 the target type. For example, warn if a @code{const char *} is cast
3935 to an ordinary @code{char *}.
3937 Also warn when making a cast which introduces a type qualifier in an
3938 unsafe way. For example, casting @code{char **} to @code{const char **}
3939 is unsafe, as in this example:
3942 /* p is char ** value. */
3943 const char **q = (const char **) p;
3944 /* Assignment of readonly string to const char * is OK. */
3946 /* Now char** pointer points to read-only memory. */
3951 @opindex Wcast-align
3952 @opindex Wno-cast-align
3953 Warn whenever a pointer is cast such that the required alignment of the
3954 target is increased. For example, warn if a @code{char *} is cast to
3955 an @code{int *} on machines where integers can only be accessed at
3956 two- or four-byte boundaries.
3958 @item -Wwrite-strings
3959 @opindex Wwrite-strings
3960 @opindex Wno-write-strings
3961 When compiling C, give string constants the type @code{const
3962 char[@var{length}]} so that copying the address of one into a
3963 non-@code{const} @code{char *} pointer will get a warning. These
3964 warnings will help you find at compile time code that can try to write
3965 into a string constant, but only if you have been very careful about
3966 using @code{const} in declarations and prototypes. Otherwise, it will
3967 just be a nuisance. This is why we did not make @option{-Wall} request
3970 When compiling C++, warn about the deprecated conversion from string
3971 literals to @code{char *}. This warning is enabled by default for C++
3976 @opindex Wno-clobbered
3977 Warn for variables that might be changed by @samp{longjmp} or
3978 @samp{vfork}. This warning is also enabled by @option{-Wextra}.
3981 @opindex Wconversion
3982 @opindex Wno-conversion
3983 Warn for implicit conversions that may alter a value. This includes
3984 conversions between real and integer, like @code{abs (x)} when
3985 @code{x} is @code{double}; conversions between signed and unsigned,
3986 like @code{unsigned ui = -1}; and conversions to smaller types, like
3987 @code{sqrtf (M_PI)}. Do not warn for explicit casts like @code{abs
3988 ((int) x)} and @code{ui = (unsigned) -1}, or if the value is not
3989 changed by the conversion like in @code{abs (2.0)}. Warnings about
3990 conversions between signed and unsigned integers can be disabled by
3991 using @option{-Wno-sign-conversion}.
3993 For C++, also warn for confusing overload resolution for user-defined
3994 conversions; and conversions that will never use a type conversion
3995 operator: conversions to @code{void}, the same type, a base class or a
3996 reference to them. Warnings about conversions between signed and
3997 unsigned integers are disabled by default in C++ unless
3998 @option{-Wsign-conversion} is explicitly enabled.
4000 @item -Wno-conversion-null @r{(C++ and Objective-C++ only)}
4001 @opindex Wconversion-null
4002 @opindex Wno-conversion-null
4003 Do not warn for conversions between @code{NULL} and non-pointer
4004 types. @option{-Wconversion-null} is enabled by default.
4007 @opindex Wempty-body
4008 @opindex Wno-empty-body
4009 Warn if an empty body occurs in an @samp{if}, @samp{else} or @samp{do
4010 while} statement. This warning is also enabled by @option{-Wextra}.
4012 @item -Wenum-compare
4013 @opindex Wenum-compare
4014 @opindex Wno-enum-compare
4015 Warn about a comparison between values of different enum types. In C++
4016 this warning is enabled by default. In C this warning is enabled by
4019 @item -Wjump-misses-init @r{(C, Objective-C only)}
4020 @opindex Wjump-misses-init
4021 @opindex Wno-jump-misses-init
4022 Warn if a @code{goto} statement or a @code{switch} statement jumps
4023 forward across the initialization of a variable, or jumps backward to a
4024 label after the variable has been initialized. This only warns about
4025 variables which are initialized when they are declared. This warning is
4026 only supported for C and Objective C; in C++ this sort of branch is an
4029 @option{-Wjump-misses-init} is included in @option{-Wc++-compat}. It
4030 can be disabled with the @option{-Wno-jump-misses-init} option.
4032 @item -Wsign-compare
4033 @opindex Wsign-compare
4034 @opindex Wno-sign-compare
4035 @cindex warning for comparison of signed and unsigned values
4036 @cindex comparison of signed and unsigned values, warning
4037 @cindex signed and unsigned values, comparison warning
4038 Warn when a comparison between signed and unsigned values could produce
4039 an incorrect result when the signed value is converted to unsigned.
4040 This warning is also enabled by @option{-Wextra}; to get the other warnings
4041 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
4043 @item -Wsign-conversion
4044 @opindex Wsign-conversion
4045 @opindex Wno-sign-conversion
4046 Warn for implicit conversions that may change the sign of an integer
4047 value, like assigning a signed integer expression to an unsigned
4048 integer variable. An explicit cast silences the warning. In C, this
4049 option is enabled also by @option{-Wconversion}.
4053 @opindex Wno-address
4054 Warn about suspicious uses of memory addresses. These include using
4055 the address of a function in a conditional expression, such as
4056 @code{void func(void); if (func)}, and comparisons against the memory
4057 address of a string literal, such as @code{if (x == "abc")}. Such
4058 uses typically indicate a programmer error: the address of a function
4059 always evaluates to true, so their use in a conditional usually
4060 indicate that the programmer forgot the parentheses in a function
4061 call; and comparisons against string literals result in unspecified
4062 behavior and are not portable in C, so they usually indicate that the
4063 programmer intended to use @code{strcmp}. This warning is enabled by
4067 @opindex Wlogical-op
4068 @opindex Wno-logical-op
4069 Warn about suspicious uses of logical operators in expressions.
4070 This includes using logical operators in contexts where a
4071 bit-wise operator is likely to be expected.
4073 @item -Waggregate-return
4074 @opindex Waggregate-return
4075 @opindex Wno-aggregate-return
4076 Warn if any functions that return structures or unions are defined or
4077 called. (In languages where you can return an array, this also elicits
4080 @item -Wno-attributes
4081 @opindex Wno-attributes
4082 @opindex Wattributes
4083 Do not warn if an unexpected @code{__attribute__} is used, such as
4084 unrecognized attributes, function attributes applied to variables,
4085 etc. This will not stop errors for incorrect use of supported
4088 @item -Wno-builtin-macro-redefined
4089 @opindex Wno-builtin-macro-redefined
4090 @opindex Wbuiltin-macro-redefined
4091 Do not warn if certain built-in macros are redefined. This suppresses
4092 warnings for redefinition of @code{__TIMESTAMP__}, @code{__TIME__},
4093 @code{__DATE__}, @code{__FILE__}, and @code{__BASE_FILE__}.
4095 @item -Wstrict-prototypes @r{(C and Objective-C only)}
4096 @opindex Wstrict-prototypes
4097 @opindex Wno-strict-prototypes
4098 Warn if a function is declared or defined without specifying the
4099 argument types. (An old-style function definition is permitted without
4100 a warning if preceded by a declaration which specifies the argument
4103 @item -Wold-style-declaration @r{(C and Objective-C only)}
4104 @opindex Wold-style-declaration
4105 @opindex Wno-old-style-declaration
4106 Warn for obsolescent usages, according to the C Standard, in a
4107 declaration. For example, warn if storage-class specifiers like
4108 @code{static} are not the first things in a declaration. This warning
4109 is also enabled by @option{-Wextra}.
4111 @item -Wold-style-definition @r{(C and Objective-C only)}
4112 @opindex Wold-style-definition
4113 @opindex Wno-old-style-definition
4114 Warn if an old-style function definition is used. A warning is given
4115 even if there is a previous prototype.
4117 @item -Wmissing-parameter-type @r{(C and Objective-C only)}
4118 @opindex Wmissing-parameter-type
4119 @opindex Wno-missing-parameter-type
4120 A function parameter is declared without a type specifier in K&R-style
4127 This warning is also enabled by @option{-Wextra}.
4129 @item -Wmissing-prototypes @r{(C and Objective-C only)}
4130 @opindex Wmissing-prototypes
4131 @opindex Wno-missing-prototypes
4132 Warn if a global function is defined without a previous prototype
4133 declaration. This warning is issued even if the definition itself
4134 provides a prototype. The aim is to detect global functions that fail
4135 to be declared in header files.
4137 @item -Wmissing-declarations
4138 @opindex Wmissing-declarations
4139 @opindex Wno-missing-declarations
4140 Warn if a global function is defined without a previous declaration.
4141 Do so even if the definition itself provides a prototype.
4142 Use this option to detect global functions that are not declared in
4143 header files. In C++, no warnings are issued for function templates,
4144 or for inline functions, or for functions in anonymous namespaces.
4146 @item -Wmissing-field-initializers
4147 @opindex Wmissing-field-initializers
4148 @opindex Wno-missing-field-initializers
4152 Warn if a structure's initializer has some fields missing. For
4153 example, the following code would cause such a warning, because
4154 @code{x.h} is implicitly zero:
4157 struct s @{ int f, g, h; @};
4158 struct s x = @{ 3, 4 @};
4161 This option does not warn about designated initializers, so the following
4162 modification would not trigger a warning:
4165 struct s @{ int f, g, h; @};
4166 struct s x = @{ .f = 3, .g = 4 @};
4169 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
4170 warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}.
4172 @item -Wmissing-format-attribute
4173 @opindex Wmissing-format-attribute
4174 @opindex Wno-missing-format-attribute
4177 Warn about function pointers which might be candidates for @code{format}
4178 attributes. Note these are only possible candidates, not absolute ones.
4179 GCC will guess that function pointers with @code{format} attributes that
4180 are used in assignment, initialization, parameter passing or return
4181 statements should have a corresponding @code{format} attribute in the
4182 resulting type. I.e.@: the left-hand side of the assignment or
4183 initialization, the type of the parameter variable, or the return type
4184 of the containing function respectively should also have a @code{format}
4185 attribute to avoid the warning.
4187 GCC will also warn about function definitions which might be
4188 candidates for @code{format} attributes. Again, these are only
4189 possible candidates. GCC will guess that @code{format} attributes
4190 might be appropriate for any function that calls a function like
4191 @code{vprintf} or @code{vscanf}, but this might not always be the
4192 case, and some functions for which @code{format} attributes are
4193 appropriate may not be detected.
4195 @item -Wno-multichar
4196 @opindex Wno-multichar
4198 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
4199 Usually they indicate a typo in the user's code, as they have
4200 implementation-defined values, and should not be used in portable code.
4202 @item -Wnormalized=<none|id|nfc|nfkc>
4203 @opindex Wnormalized=
4206 @cindex character set, input normalization
4207 In ISO C and ISO C++, two identifiers are different if they are
4208 different sequences of characters. However, sometimes when characters
4209 outside the basic ASCII character set are used, you can have two
4210 different character sequences that look the same. To avoid confusion,
4211 the ISO 10646 standard sets out some @dfn{normalization rules} which
4212 when applied ensure that two sequences that look the same are turned into
4213 the same sequence. GCC can warn you if you are using identifiers which
4214 have not been normalized; this option controls that warning.
4216 There are four levels of warning that GCC supports. The default is
4217 @option{-Wnormalized=nfc}, which warns about any identifier which is
4218 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
4219 recommended form for most uses.
4221 Unfortunately, there are some characters which ISO C and ISO C++ allow
4222 in identifiers that when turned into NFC aren't allowable as
4223 identifiers. That is, there's no way to use these symbols in portable
4224 ISO C or C++ and have all your identifiers in NFC@.
4225 @option{-Wnormalized=id} suppresses the warning for these characters.
4226 It is hoped that future versions of the standards involved will correct
4227 this, which is why this option is not the default.
4229 You can switch the warning off for all characters by writing
4230 @option{-Wnormalized=none}. You would only want to do this if you
4231 were using some other normalization scheme (like ``D''), because
4232 otherwise you can easily create bugs that are literally impossible to see.
4234 Some characters in ISO 10646 have distinct meanings but look identical
4235 in some fonts or display methodologies, especially once formatting has
4236 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
4237 LETTER N'', will display just like a regular @code{n} which has been
4238 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
4239 normalization scheme to convert all these into a standard form as
4240 well, and GCC will warn if your code is not in NFKC if you use
4241 @option{-Wnormalized=nfkc}. This warning is comparable to warning
4242 about every identifier that contains the letter O because it might be
4243 confused with the digit 0, and so is not the default, but may be
4244 useful as a local coding convention if the programming environment is
4245 unable to be fixed to display these characters distinctly.
4247 @item -Wno-deprecated
4248 @opindex Wno-deprecated
4249 @opindex Wdeprecated
4250 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
4252 @item -Wno-deprecated-declarations
4253 @opindex Wno-deprecated-declarations
4254 @opindex Wdeprecated-declarations
4255 Do not warn about uses of functions (@pxref{Function Attributes}),
4256 variables (@pxref{Variable Attributes}), and types (@pxref{Type
4257 Attributes}) marked as deprecated by using the @code{deprecated}
4261 @opindex Wno-overflow
4263 Do not warn about compile-time overflow in constant expressions.
4265 @item -Woverride-init @r{(C and Objective-C only)}
4266 @opindex Woverride-init
4267 @opindex Wno-override-init
4271 Warn if an initialized field without side effects is overridden when
4272 using designated initializers (@pxref{Designated Inits, , Designated
4275 This warning is included in @option{-Wextra}. To get other
4276 @option{-Wextra} warnings without this one, use @samp{-Wextra
4277 -Wno-override-init}.
4282 Warn if a structure is given the packed attribute, but the packed
4283 attribute has no effect on the layout or size of the structure.
4284 Such structures may be mis-aligned for little benefit. For
4285 instance, in this code, the variable @code{f.x} in @code{struct bar}
4286 will be misaligned even though @code{struct bar} does not itself
4287 have the packed attribute:
4294 @} __attribute__((packed));
4302 @item -Wpacked-bitfield-compat
4303 @opindex Wpacked-bitfield-compat
4304 @opindex Wno-packed-bitfield-compat
4305 The 4.1, 4.2 and 4.3 series of GCC ignore the @code{packed} attribute
4306 on bit-fields of type @code{char}. This has been fixed in GCC 4.4 but
4307 the change can lead to differences in the structure layout. GCC
4308 informs you when the offset of such a field has changed in GCC 4.4.
4309 For example there is no longer a 4-bit padding between field @code{a}
4310 and @code{b} in this structure:
4317 @} __attribute__ ((packed));
4320 This warning is enabled by default. Use
4321 @option{-Wno-packed-bitfield-compat} to disable this warning.
4326 Warn if padding is included in a structure, either to align an element
4327 of the structure or to align the whole structure. Sometimes when this
4328 happens it is possible to rearrange the fields of the structure to
4329 reduce the padding and so make the structure smaller.
4331 @item -Wredundant-decls
4332 @opindex Wredundant-decls
4333 @opindex Wno-redundant-decls
4334 Warn if anything is declared more than once in the same scope, even in
4335 cases where multiple declaration is valid and changes nothing.
4337 @item -Wnested-externs @r{(C and Objective-C only)}
4338 @opindex Wnested-externs
4339 @opindex Wno-nested-externs
4340 Warn if an @code{extern} declaration is encountered within a function.
4345 Warn if a function can not be inlined and it was declared as inline.
4346 Even with this option, the compiler will not warn about failures to
4347 inline functions declared in system headers.
4349 The compiler uses a variety of heuristics to determine whether or not
4350 to inline a function. For example, the compiler takes into account
4351 the size of the function being inlined and the amount of inlining
4352 that has already been done in the current function. Therefore,
4353 seemingly insignificant changes in the source program can cause the
4354 warnings produced by @option{-Winline} to appear or disappear.
4356 @item -Wno-invalid-offsetof @r{(C++ and Objective-C++ only)}
4357 @opindex Wno-invalid-offsetof
4358 @opindex Winvalid-offsetof
4359 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
4360 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
4361 to a non-POD type is undefined. In existing C++ implementations,
4362 however, @samp{offsetof} typically gives meaningful results even when
4363 applied to certain kinds of non-POD types. (Such as a simple
4364 @samp{struct} that fails to be a POD type only by virtue of having a
4365 constructor.) This flag is for users who are aware that they are
4366 writing nonportable code and who have deliberately chosen to ignore the
4369 The restrictions on @samp{offsetof} may be relaxed in a future version
4370 of the C++ standard.
4372 @item -Wno-int-to-pointer-cast
4373 @opindex Wno-int-to-pointer-cast
4374 @opindex Wint-to-pointer-cast
4375 Suppress warnings from casts to pointer type of an integer of a
4376 different size. In C++, casting to a pointer type of smaller size is
4377 an error. @option{Wint-to-pointer-cast} is enabled by default.
4380 @item -Wno-pointer-to-int-cast @r{(C and Objective-C only)}
4381 @opindex Wno-pointer-to-int-cast
4382 @opindex Wpointer-to-int-cast
4383 Suppress warnings from casts from a pointer to an integer type of a
4387 @opindex Winvalid-pch
4388 @opindex Wno-invalid-pch
4389 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
4390 the search path but can't be used.
4394 @opindex Wno-long-long
4395 Warn if @samp{long long} type is used. This is enabled by either
4396 @option{-pedantic} or @option{-Wtraditional} in ISO C90 and C++98
4397 modes. To inhibit the warning messages, use @option{-Wno-long-long}.
4399 @item -Wvariadic-macros
4400 @opindex Wvariadic-macros
4401 @opindex Wno-variadic-macros
4402 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
4403 alternate syntax when in pedantic ISO C99 mode. This is default.
4404 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
4409 Warn if variable length array is used in the code.
4410 @option{-Wno-vla} will prevent the @option{-pedantic} warning of
4411 the variable length array.
4413 @item -Wvolatile-register-var
4414 @opindex Wvolatile-register-var
4415 @opindex Wno-volatile-register-var
4416 Warn if a register variable is declared volatile. The volatile
4417 modifier does not inhibit all optimizations that may eliminate reads
4418 and/or writes to register variables. This warning is enabled by
4421 @item -Wdisabled-optimization
4422 @opindex Wdisabled-optimization
4423 @opindex Wno-disabled-optimization
4424 Warn if a requested optimization pass is disabled. This warning does
4425 not generally indicate that there is anything wrong with your code; it
4426 merely indicates that GCC's optimizers were unable to handle the code
4427 effectively. Often, the problem is that your code is too big or too
4428 complex; GCC will refuse to optimize programs when the optimization
4429 itself is likely to take inordinate amounts of time.
4431 @item -Wpointer-sign @r{(C and Objective-C only)}
4432 @opindex Wpointer-sign
4433 @opindex Wno-pointer-sign
4434 Warn for pointer argument passing or assignment with different signedness.
4435 This option is only supported for C and Objective-C@. It is implied by
4436 @option{-Wall} and by @option{-pedantic}, which can be disabled with
4437 @option{-Wno-pointer-sign}.
4439 @item -Wstack-protector
4440 @opindex Wstack-protector
4441 @opindex Wno-stack-protector
4442 This option is only active when @option{-fstack-protector} is active. It
4443 warns about functions that will not be protected against stack smashing.
4446 @opindex Wno-mudflap
4447 Suppress warnings about constructs that cannot be instrumented by
4450 @item -Woverlength-strings
4451 @opindex Woverlength-strings
4452 @opindex Wno-overlength-strings
4453 Warn about string constants which are longer than the ``minimum
4454 maximum'' length specified in the C standard. Modern compilers
4455 generally allow string constants which are much longer than the
4456 standard's minimum limit, but very portable programs should avoid
4457 using longer strings.
4459 The limit applies @emph{after} string constant concatenation, and does
4460 not count the trailing NUL@. In C90, the limit was 509 characters; in
4461 C99, it was raised to 4095. C++98 does not specify a normative
4462 minimum maximum, so we do not diagnose overlength strings in C++@.
4464 This option is implied by @option{-pedantic}, and can be disabled with
4465 @option{-Wno-overlength-strings}.
4467 @item -Wunsuffixed-float-constants @r{(C and Objective-C only)}
4468 @opindex Wunsuffixed-float-constants
4470 GCC will issue a warning for any floating constant that does not have
4471 a suffix. When used together with @option{-Wsystem-headers} it will
4472 warn about such constants in system header files. This can be useful
4473 when preparing code to use with the @code{FLOAT_CONST_DECIMAL64} pragma
4474 from the decimal floating-point extension to C99.
4477 @node Debugging Options
4478 @section Options for Debugging Your Program or GCC
4479 @cindex options, debugging
4480 @cindex debugging information options
4482 GCC has various special options that are used for debugging
4483 either your program or GCC:
4488 Produce debugging information in the operating system's native format
4489 (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging
4492 On most systems that use stabs format, @option{-g} enables use of extra
4493 debugging information that only GDB can use; this extra information
4494 makes debugging work better in GDB but will probably make other debuggers
4496 refuse to read the program. If you want to control for certain whether
4497 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
4498 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
4500 GCC allows you to use @option{-g} with
4501 @option{-O}. The shortcuts taken by optimized code may occasionally
4502 produce surprising results: some variables you declared may not exist
4503 at all; flow of control may briefly move where you did not expect it;
4504 some statements may not be executed because they compute constant
4505 results or their values were already at hand; some statements may
4506 execute in different places because they were moved out of loops.
4508 Nevertheless it proves possible to debug optimized output. This makes
4509 it reasonable to use the optimizer for programs that might have bugs.
4511 The following options are useful when GCC is generated with the
4512 capability for more than one debugging format.
4516 Produce debugging information for use by GDB@. This means to use the
4517 most expressive format available (DWARF 2, stabs, or the native format
4518 if neither of those are supported), including GDB extensions if at all
4523 Produce debugging information in stabs format (if that is supported),
4524 without GDB extensions. This is the format used by DBX on most BSD
4525 systems. On MIPS, Alpha and System V Release 4 systems this option
4526 produces stabs debugging output which is not understood by DBX or SDB@.
4527 On System V Release 4 systems this option requires the GNU assembler.
4529 @item -feliminate-unused-debug-symbols
4530 @opindex feliminate-unused-debug-symbols
4531 Produce debugging information in stabs format (if that is supported),
4532 for only symbols that are actually used.
4534 @item -femit-class-debug-always
4535 Instead of emitting debugging information for a C++ class in only one
4536 object file, emit it in all object files using the class. This option
4537 should be used only with debuggers that are unable to handle the way GCC
4538 normally emits debugging information for classes because using this
4539 option will increase the size of debugging information by as much as a
4544 Produce debugging information in stabs format (if that is supported),
4545 using GNU extensions understood only by the GNU debugger (GDB)@. The
4546 use of these extensions is likely to make other debuggers crash or
4547 refuse to read the program.
4551 Produce debugging information in COFF format (if that is supported).
4552 This is the format used by SDB on most System V systems prior to
4557 Produce debugging information in XCOFF format (if that is supported).
4558 This is the format used by the DBX debugger on IBM RS/6000 systems.
4562 Produce debugging information in XCOFF format (if that is supported),
4563 using GNU extensions understood only by the GNU debugger (GDB)@. The
4564 use of these extensions is likely to make other debuggers crash or
4565 refuse to read the program, and may cause assemblers other than the GNU
4566 assembler (GAS) to fail with an error.
4568 @item -gdwarf-@var{version}
4569 @opindex gdwarf-@var{version}
4570 Produce debugging information in DWARF format (if that is
4571 supported). This is the format used by DBX on IRIX 6. The value
4572 of @var{version} may be either 2, 3 or 4; the default version is 2.
4574 Note that with DWARF version 2 some ports require, and will always
4575 use, some non-conflicting DWARF 3 extensions in the unwind tables.
4577 Version 4 may require GDB 7.0 and @option{-fvar-tracking-assignments}
4578 for maximum benefit.
4580 @item -gstrict-dwarf
4581 @opindex gstrict-dwarf
4582 Disallow using extensions of later DWARF standard version than selected
4583 with @option{-gdwarf-@var{version}}. On most targets using non-conflicting
4584 DWARF extensions from later standard versions is allowed.
4586 @item -gno-strict-dwarf
4587 @opindex gno-strict-dwarf
4588 Allow using extensions of later DWARF standard version than selected with
4589 @option{-gdwarf-@var{version}}.
4593 Produce debugging information in VMS debug format (if that is
4594 supported). This is the format used by DEBUG on VMS systems.
4597 @itemx -ggdb@var{level}
4598 @itemx -gstabs@var{level}
4599 @itemx -gcoff@var{level}
4600 @itemx -gxcoff@var{level}
4601 @itemx -gvms@var{level}
4602 Request debugging information and also use @var{level} to specify how
4603 much information. The default level is 2.
4605 Level 0 produces no debug information at all. Thus, @option{-g0} negates
4608 Level 1 produces minimal information, enough for making backtraces in
4609 parts of the program that you don't plan to debug. This includes
4610 descriptions of functions and external variables, but no information
4611 about local variables and no line numbers.
4613 Level 3 includes extra information, such as all the macro definitions
4614 present in the program. Some debuggers support macro expansion when
4615 you use @option{-g3}.
4617 @option{-gdwarf-2} does not accept a concatenated debug level, because
4618 GCC used to support an option @option{-gdwarf} that meant to generate
4619 debug information in version 1 of the DWARF format (which is very
4620 different from version 2), and it would have been too confusing. That
4621 debug format is long obsolete, but the option cannot be changed now.
4622 Instead use an additional @option{-g@var{level}} option to change the
4623 debug level for DWARF.
4627 Turn off generation of debug info, if leaving out this option would have
4628 generated it, or turn it on at level 2 otherwise. The position of this
4629 argument in the command line does not matter, it takes effect after all
4630 other options are processed, and it does so only once, no matter how
4631 many times it is given. This is mainly intended to be used with
4632 @option{-fcompare-debug}.
4634 @item -fdump-final-insns@r{[}=@var{file}@r{]}
4635 @opindex fdump-final-insns
4636 Dump the final internal representation (RTL) to @var{file}. If the
4637 optional argument is omitted (or if @var{file} is @code{.}), the name
4638 of the dump file will be determined by appending @code{.gkd} to the
4639 compilation output file name.
4641 @item -fcompare-debug@r{[}=@var{opts}@r{]}
4642 @opindex fcompare-debug
4643 @opindex fno-compare-debug
4644 If no error occurs during compilation, run the compiler a second time,
4645 adding @var{opts} and @option{-fcompare-debug-second} to the arguments
4646 passed to the second compilation. Dump the final internal
4647 representation in both compilations, and print an error if they differ.
4649 If the equal sign is omitted, the default @option{-gtoggle} is used.
4651 The environment variable @env{GCC_COMPARE_DEBUG}, if defined, non-empty
4652 and nonzero, implicitly enables @option{-fcompare-debug}. If
4653 @env{GCC_COMPARE_DEBUG} is defined to a string starting with a dash,
4654 then it is used for @var{opts}, otherwise the default @option{-gtoggle}
4657 @option{-fcompare-debug=}, with the equal sign but without @var{opts},
4658 is equivalent to @option{-fno-compare-debug}, which disables the dumping
4659 of the final representation and the second compilation, preventing even
4660 @env{GCC_COMPARE_DEBUG} from taking effect.
4662 To verify full coverage during @option{-fcompare-debug} testing, set
4663 @env{GCC_COMPARE_DEBUG} to say @samp{-fcompare-debug-not-overridden},
4664 which GCC will reject as an invalid option in any actual compilation
4665 (rather than preprocessing, assembly or linking). To get just a
4666 warning, setting @env{GCC_COMPARE_DEBUG} to @samp{-w%n-fcompare-debug
4667 not overridden} will do.
4669 @item -fcompare-debug-second
4670 @opindex fcompare-debug-second
4671 This option is implicitly passed to the compiler for the second
4672 compilation requested by @option{-fcompare-debug}, along with options to
4673 silence warnings, and omitting other options that would cause
4674 side-effect compiler outputs to files or to the standard output. Dump
4675 files and preserved temporary files are renamed so as to contain the
4676 @code{.gk} additional extension during the second compilation, to avoid
4677 overwriting those generated by the first.
4679 When this option is passed to the compiler driver, it causes the
4680 @emph{first} compilation to be skipped, which makes it useful for little
4681 other than debugging the compiler proper.
4683 @item -feliminate-dwarf2-dups
4684 @opindex feliminate-dwarf2-dups
4685 Compress DWARF2 debugging information by eliminating duplicated
4686 information about each symbol. This option only makes sense when
4687 generating DWARF2 debugging information with @option{-gdwarf-2}.
4689 @item -femit-struct-debug-baseonly
4690 Emit debug information for struct-like types
4691 only when the base name of the compilation source file
4692 matches the base name of file in which the struct was defined.
4694 This option substantially reduces the size of debugging information,
4695 but at significant potential loss in type information to the debugger.
4696 See @option{-femit-struct-debug-reduced} for a less aggressive option.
4697 See @option{-femit-struct-debug-detailed} for more detailed control.
4699 This option works only with DWARF 2.
4701 @item -femit-struct-debug-reduced
4702 Emit debug information for struct-like types
4703 only when the base name of the compilation source file
4704 matches the base name of file in which the type was defined,
4705 unless the struct is a template or defined in a system header.
4707 This option significantly reduces the size of debugging information,
4708 with some potential loss in type information to the debugger.
4709 See @option{-femit-struct-debug-baseonly} for a more aggressive option.
4710 See @option{-femit-struct-debug-detailed} for more detailed control.
4712 This option works only with DWARF 2.
4714 @item -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]}
4715 Specify the struct-like types
4716 for which the compiler will generate debug information.
4717 The intent is to reduce duplicate struct debug information
4718 between different object files within the same program.
4720 This option is a detailed version of
4721 @option{-femit-struct-debug-reduced} and @option{-femit-struct-debug-baseonly},
4722 which will serve for most needs.
4724 A specification has the syntax
4725 [@samp{dir:}|@samp{ind:}][@samp{ord:}|@samp{gen:}](@samp{any}|@samp{sys}|@samp{base}|@samp{none})
4727 The optional first word limits the specification to
4728 structs that are used directly (@samp{dir:}) or used indirectly (@samp{ind:}).
4729 A struct type is used directly when it is the type of a variable, member.
4730 Indirect uses arise through pointers to structs.
4731 That is, when use of an incomplete struct would be legal, the use is indirect.
4733 @samp{struct one direct; struct two * indirect;}.
4735 The optional second word limits the specification to
4736 ordinary structs (@samp{ord:}) or generic structs (@samp{gen:}).
4737 Generic structs are a bit complicated to explain.
4738 For C++, these are non-explicit specializations of template classes,
4739 or non-template classes within the above.
4740 Other programming languages have generics,
4741 but @samp{-femit-struct-debug-detailed} does not yet implement them.
4743 The third word specifies the source files for those
4744 structs for which the compiler will emit debug information.
4745 The values @samp{none} and @samp{any} have the normal meaning.
4746 The value @samp{base} means that
4747 the base of name of the file in which the type declaration appears
4748 must match the base of the name of the main compilation file.
4749 In practice, this means that
4750 types declared in @file{foo.c} and @file{foo.h} will have debug information,
4751 but types declared in other header will not.
4752 The value @samp{sys} means those types satisfying @samp{base}
4753 or declared in system or compiler headers.
4755 You may need to experiment to determine the best settings for your application.
4757 The default is @samp{-femit-struct-debug-detailed=all}.
4759 This option works only with DWARF 2.
4761 @item -fenable-icf-debug
4762 @opindex fenable-icf-debug
4763 Generate additional debug information to support identical code folding (ICF).
4764 This option only works with DWARF version 2 or higher.
4766 @item -fno-merge-debug-strings
4767 @opindex fmerge-debug-strings
4768 @opindex fno-merge-debug-strings
4769 Direct the linker to not merge together strings in the debugging
4770 information which are identical in different object files. Merging is
4771 not supported by all assemblers or linkers. Merging decreases the size
4772 of the debug information in the output file at the cost of increasing
4773 link processing time. Merging is enabled by default.
4775 @item -fdebug-prefix-map=@var{old}=@var{new}
4776 @opindex fdebug-prefix-map
4777 When compiling files in directory @file{@var{old}}, record debugging
4778 information describing them as in @file{@var{new}} instead.
4780 @item -fno-dwarf2-cfi-asm
4781 @opindex fdwarf2-cfi-asm
4782 @opindex fno-dwarf2-cfi-asm
4783 Emit DWARF 2 unwind info as compiler generated @code{.eh_frame} section
4784 instead of using GAS @code{.cfi_*} directives.
4786 @cindex @command{prof}
4789 Generate extra code to write profile information suitable for the
4790 analysis program @command{prof}. You must use this option when compiling
4791 the source files you want data about, and you must also use it when
4794 @cindex @command{gprof}
4797 Generate extra code to write profile information suitable for the
4798 analysis program @command{gprof}. You must use this option when compiling
4799 the source files you want data about, and you must also use it when
4804 Makes the compiler print out each function name as it is compiled, and
4805 print some statistics about each pass when it finishes.
4808 @opindex ftime-report
4809 Makes the compiler print some statistics about the time consumed by each
4810 pass when it finishes.
4813 @opindex fmem-report
4814 Makes the compiler print some statistics about permanent memory
4815 allocation when it finishes.
4817 @item -fpre-ipa-mem-report
4818 @opindex fpre-ipa-mem-report
4819 @item -fpost-ipa-mem-report
4820 @opindex fpost-ipa-mem-report
4821 Makes the compiler print some statistics about permanent memory
4822 allocation before or after interprocedural optimization.
4824 @item -fprofile-arcs
4825 @opindex fprofile-arcs
4826 Add code so that program flow @dfn{arcs} are instrumented. During
4827 execution the program records how many times each branch and call is
4828 executed and how many times it is taken or returns. When the compiled
4829 program exits it saves this data to a file called
4830 @file{@var{auxname}.gcda} for each source file. The data may be used for
4831 profile-directed optimizations (@option{-fbranch-probabilities}), or for
4832 test coverage analysis (@option{-ftest-coverage}). Each object file's
4833 @var{auxname} is generated from the name of the output file, if
4834 explicitly specified and it is not the final executable, otherwise it is
4835 the basename of the source file. In both cases any suffix is removed
4836 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
4837 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
4838 @xref{Cross-profiling}.
4840 @cindex @command{gcov}
4844 This option is used to compile and link code instrumented for coverage
4845 analysis. The option is a synonym for @option{-fprofile-arcs}
4846 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
4847 linking). See the documentation for those options for more details.
4852 Compile the source files with @option{-fprofile-arcs} plus optimization
4853 and code generation options. For test coverage analysis, use the
4854 additional @option{-ftest-coverage} option. You do not need to profile
4855 every source file in a program.
4858 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
4859 (the latter implies the former).
4862 Run the program on a representative workload to generate the arc profile
4863 information. This may be repeated any number of times. You can run
4864 concurrent instances of your program, and provided that the file system
4865 supports locking, the data files will be correctly updated. Also
4866 @code{fork} calls are detected and correctly handled (double counting
4870 For profile-directed optimizations, compile the source files again with
4871 the same optimization and code generation options plus
4872 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
4873 Control Optimization}).
4876 For test coverage analysis, use @command{gcov} to produce human readable
4877 information from the @file{.gcno} and @file{.gcda} files. Refer to the
4878 @command{gcov} documentation for further information.
4882 With @option{-fprofile-arcs}, for each function of your program GCC
4883 creates a program flow graph, then finds a spanning tree for the graph.
4884 Only arcs that are not on the spanning tree have to be instrumented: the
4885 compiler adds code to count the number of times that these arcs are
4886 executed. When an arc is the only exit or only entrance to a block, the
4887 instrumentation code can be added to the block; otherwise, a new basic
4888 block must be created to hold the instrumentation code.
4891 @item -ftest-coverage
4892 @opindex ftest-coverage
4893 Produce a notes file that the @command{gcov} code-coverage utility
4894 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
4895 show program coverage. Each source file's note file is called
4896 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
4897 above for a description of @var{auxname} and instructions on how to
4898 generate test coverage data. Coverage data will match the source files
4899 more closely, if you do not optimize.
4901 @item -fdbg-cnt-list
4902 @opindex fdbg-cnt-list
4903 Print the name and the counter upperbound for all debug counters.
4905 @item -fdbg-cnt=@var{counter-value-list}
4907 Set the internal debug counter upperbound. @var{counter-value-list}
4908 is a comma-separated list of @var{name}:@var{value} pairs
4909 which sets the upperbound of each debug counter @var{name} to @var{value}.
4910 All debug counters have the initial upperbound of @var{UINT_MAX},
4911 thus dbg_cnt() returns true always unless the upperbound is set by this option.
4912 e.g. With -fdbg-cnt=dce:10,tail_call:0
4913 dbg_cnt(dce) will return true only for first 10 invocations
4914 and dbg_cnt(tail_call) will return false always.
4916 @item -d@var{letters}
4917 @itemx -fdump-rtl-@var{pass}
4919 Says to make debugging dumps during compilation at times specified by
4920 @var{letters}. This is used for debugging the RTL-based passes of the
4921 compiler. The file names for most of the dumps are made by appending
4922 a pass number and a word to the @var{dumpname}, and the files are
4923 created in the directory of the output file. @var{dumpname} is
4924 generated from the name of the output file, if explicitly specified
4925 and it is not an executable, otherwise it is the basename of the
4926 source file. These switches may have different effects when
4927 @option{-E} is used for preprocessing.
4929 Debug dumps can be enabled with a @option{-fdump-rtl} switch or some
4930 @option{-d} option @var{letters}. Here are the possible
4931 letters for use in @var{pass} and @var{letters}, and their meanings:
4935 @item -fdump-rtl-alignments
4936 @opindex fdump-rtl-alignments
4937 Dump after branch alignments have been computed.
4939 @item -fdump-rtl-asmcons
4940 @opindex fdump-rtl-asmcons
4941 Dump after fixing rtl statements that have unsatisfied in/out constraints.
4943 @item -fdump-rtl-auto_inc_dec
4944 @opindex fdump-rtl-auto_inc_dec
4945 Dump after auto-inc-dec discovery. This pass is only run on
4946 architectures that have auto inc or auto dec instructions.
4948 @item -fdump-rtl-barriers
4949 @opindex fdump-rtl-barriers
4950 Dump after cleaning up the barrier instructions.
4952 @item -fdump-rtl-bbpart
4953 @opindex fdump-rtl-bbpart
4954 Dump after partitioning hot and cold basic blocks.
4956 @item -fdump-rtl-bbro
4957 @opindex fdump-rtl-bbro
4958 Dump after block reordering.
4960 @item -fdump-rtl-btl1
4961 @itemx -fdump-rtl-btl2
4962 @opindex fdump-rtl-btl2
4963 @opindex fdump-rtl-btl2
4964 @option{-fdump-rtl-btl1} and @option{-fdump-rtl-btl2} enable dumping
4965 after the two branch
4966 target load optimization passes.
4968 @item -fdump-rtl-bypass
4969 @opindex fdump-rtl-bypass
4970 Dump after jump bypassing and control flow optimizations.
4972 @item -fdump-rtl-combine
4973 @opindex fdump-rtl-combine
4974 Dump after the RTL instruction combination pass.
4976 @item -fdump-rtl-compgotos
4977 @opindex fdump-rtl-compgotos
4978 Dump after duplicating the computed gotos.
4980 @item -fdump-rtl-ce1
4981 @itemx -fdump-rtl-ce2
4982 @itemx -fdump-rtl-ce3
4983 @opindex fdump-rtl-ce1
4984 @opindex fdump-rtl-ce2
4985 @opindex fdump-rtl-ce3
4986 @option{-fdump-rtl-ce1}, @option{-fdump-rtl-ce2}, and
4987 @option{-fdump-rtl-ce3} enable dumping after the three
4988 if conversion passes.
4990 @itemx -fdump-rtl-cprop_hardreg
4991 @opindex fdump-rtl-cprop_hardreg
4992 Dump after hard register copy propagation.
4994 @itemx -fdump-rtl-csa
4995 @opindex fdump-rtl-csa
4996 Dump after combining stack adjustments.
4998 @item -fdump-rtl-cse1
4999 @itemx -fdump-rtl-cse2
5000 @opindex fdump-rtl-cse1
5001 @opindex fdump-rtl-cse2
5002 @option{-fdump-rtl-cse1} and @option{-fdump-rtl-cse2} enable dumping after
5003 the two common sub-expression elimination passes.
5005 @itemx -fdump-rtl-dce
5006 @opindex fdump-rtl-dce
5007 Dump after the standalone dead code elimination passes.
5009 @itemx -fdump-rtl-dbr
5010 @opindex fdump-rtl-dbr
5011 Dump after delayed branch scheduling.
5013 @item -fdump-rtl-dce1
5014 @itemx -fdump-rtl-dce2
5015 @opindex fdump-rtl-dce1
5016 @opindex fdump-rtl-dce2
5017 @option{-fdump-rtl-dce1} and @option{-fdump-rtl-dce2} enable dumping after
5018 the two dead store elimination passes.
5021 @opindex fdump-rtl-eh
5022 Dump after finalization of EH handling code.
5024 @item -fdump-rtl-eh_ranges
5025 @opindex fdump-rtl-eh_ranges
5026 Dump after conversion of EH handling range regions.
5028 @item -fdump-rtl-expand
5029 @opindex fdump-rtl-expand
5030 Dump after RTL generation.
5032 @item -fdump-rtl-fwprop1
5033 @itemx -fdump-rtl-fwprop2
5034 @opindex fdump-rtl-fwprop1
5035 @opindex fdump-rtl-fwprop2
5036 @option{-fdump-rtl-fwprop1} and @option{-fdump-rtl-fwprop2} enable
5037 dumping after the two forward propagation passes.
5039 @item -fdump-rtl-gcse1
5040 @itemx -fdump-rtl-gcse2
5041 @opindex fdump-rtl-gcse1
5042 @opindex fdump-rtl-gcse2
5043 @option{-fdump-rtl-gcse1} and @option{-fdump-rtl-gcse2} enable dumping
5044 after global common subexpression elimination.
5046 @item -fdump-rtl-init-regs
5047 @opindex fdump-rtl-init-regs
5048 Dump after the initialization of the registers.
5050 @item -fdump-rtl-initvals
5051 @opindex fdump-rtl-initvals
5052 Dump after the computation of the initial value sets.
5054 @itemx -fdump-rtl-into_cfglayout
5055 @opindex fdump-rtl-into_cfglayout
5056 Dump after converting to cfglayout mode.
5058 @item -fdump-rtl-ira
5059 @opindex fdump-rtl-ira
5060 Dump after iterated register allocation.
5062 @item -fdump-rtl-jump
5063 @opindex fdump-rtl-jump
5064 Dump after the second jump optimization.
5066 @item -fdump-rtl-loop2
5067 @opindex fdump-rtl-loop2
5068 @option{-fdump-rtl-loop2} enables dumping after the rtl
5069 loop optimization passes.
5071 @item -fdump-rtl-mach
5072 @opindex fdump-rtl-mach
5073 Dump after performing the machine dependent reorganization pass, if that
5076 @item -fdump-rtl-mode_sw
5077 @opindex fdump-rtl-mode_sw
5078 Dump after removing redundant mode switches.
5080 @item -fdump-rtl-rnreg
5081 @opindex fdump-rtl-rnreg
5082 Dump after register renumbering.
5084 @itemx -fdump-rtl-outof_cfglayout
5085 @opindex fdump-rtl-outof_cfglayout
5086 Dump after converting from cfglayout mode.
5088 @item -fdump-rtl-peephole2
5089 @opindex fdump-rtl-peephole2
5090 Dump after the peephole pass.
5092 @item -fdump-rtl-postreload
5093 @opindex fdump-rtl-postreload
5094 Dump after post-reload optimizations.
5096 @itemx -fdump-rtl-pro_and_epilogue
5097 @opindex fdump-rtl-pro_and_epilogue
5098 Dump after generating the function pro and epilogues.
5100 @item -fdump-rtl-regmove
5101 @opindex fdump-rtl-regmove
5102 Dump after the register move pass.
5104 @item -fdump-rtl-sched1
5105 @itemx -fdump-rtl-sched2
5106 @opindex fdump-rtl-sched1
5107 @opindex fdump-rtl-sched2
5108 @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2} enable dumping
5109 after the basic block scheduling passes.
5111 @item -fdump-rtl-see
5112 @opindex fdump-rtl-see
5113 Dump after sign extension elimination.
5115 @item -fdump-rtl-seqabstr
5116 @opindex fdump-rtl-seqabstr
5117 Dump after common sequence discovery.
5119 @item -fdump-rtl-shorten
5120 @opindex fdump-rtl-shorten
5121 Dump after shortening branches.
5123 @item -fdump-rtl-sibling
5124 @opindex fdump-rtl-sibling
5125 Dump after sibling call optimizations.
5127 @item -fdump-rtl-split1
5128 @itemx -fdump-rtl-split2
5129 @itemx -fdump-rtl-split3
5130 @itemx -fdump-rtl-split4
5131 @itemx -fdump-rtl-split5
5132 @opindex fdump-rtl-split1
5133 @opindex fdump-rtl-split2
5134 @opindex fdump-rtl-split3
5135 @opindex fdump-rtl-split4
5136 @opindex fdump-rtl-split5
5137 @option{-fdump-rtl-split1}, @option{-fdump-rtl-split2},
5138 @option{-fdump-rtl-split3}, @option{-fdump-rtl-split4} and
5139 @option{-fdump-rtl-split5} enable dumping after five rounds of
5140 instruction splitting.
5142 @item -fdump-rtl-sms
5143 @opindex fdump-rtl-sms
5144 Dump after modulo scheduling. This pass is only run on some
5147 @item -fdump-rtl-stack
5148 @opindex fdump-rtl-stack
5149 Dump after conversion from GCC's "flat register file" registers to the
5150 x87's stack-like registers. This pass is only run on x86 variants.
5152 @item -fdump-rtl-subreg1
5153 @itemx -fdump-rtl-subreg2
5154 @opindex fdump-rtl-subreg1
5155 @opindex fdump-rtl-subreg2
5156 @option{-fdump-rtl-subreg1} and @option{-fdump-rtl-subreg2} enable dumping after
5157 the two subreg expansion passes.
5159 @item -fdump-rtl-unshare
5160 @opindex fdump-rtl-unshare
5161 Dump after all rtl has been unshared.
5163 @item -fdump-rtl-vartrack
5164 @opindex fdump-rtl-vartrack
5165 Dump after variable tracking.
5167 @item -fdump-rtl-vregs
5168 @opindex fdump-rtl-vregs
5169 Dump after converting virtual registers to hard registers.
5171 @item -fdump-rtl-web
5172 @opindex fdump-rtl-web
5173 Dump after live range splitting.
5175 @item -fdump-rtl-regclass
5176 @itemx -fdump-rtl-subregs_of_mode_init
5177 @itemx -fdump-rtl-subregs_of_mode_finish
5178 @itemx -fdump-rtl-dfinit
5179 @itemx -fdump-rtl-dfinish
5180 @opindex fdump-rtl-regclass
5181 @opindex fdump-rtl-subregs_of_mode_init
5182 @opindex fdump-rtl-subregs_of_mode_finish
5183 @opindex fdump-rtl-dfinit
5184 @opindex fdump-rtl-dfinish
5185 These dumps are defined but always produce empty files.
5187 @item -fdump-rtl-all
5188 @opindex fdump-rtl-all
5189 Produce all the dumps listed above.
5193 Annotate the assembler output with miscellaneous debugging information.
5197 Dump all macro definitions, at the end of preprocessing, in addition to
5202 Produce a core dump whenever an error occurs.
5206 Print statistics on memory usage, at the end of the run, to
5211 Annotate the assembler output with a comment indicating which
5212 pattern and alternative was used. The length of each instruction is
5217 Dump the RTL in the assembler output as a comment before each instruction.
5218 Also turns on @option{-dp} annotation.
5222 For each of the other indicated dump files (@option{-fdump-rtl-@var{pass}}),
5223 dump a representation of the control flow graph suitable for viewing with VCG
5224 to @file{@var{file}.@var{pass}.vcg}.
5228 Just generate RTL for a function instead of compiling it. Usually used
5229 with @option{-fdump-rtl-expand}.
5233 Dump debugging information during parsing, to standard error.
5237 @opindex fdump-noaddr
5238 When doing debugging dumps, suppress address output. This makes it more
5239 feasible to use diff on debugging dumps for compiler invocations with
5240 different compiler binaries and/or different
5241 text / bss / data / heap / stack / dso start locations.
5243 @item -fdump-unnumbered
5244 @opindex fdump-unnumbered
5245 When doing debugging dumps, suppress instruction numbers and address output.
5246 This makes it more feasible to use diff on debugging dumps for compiler
5247 invocations with different options, in particular with and without
5250 @item -fdump-unnumbered-links
5251 @opindex fdump-unnumbered-links
5252 When doing debugging dumps (see @option{-d} option above), suppress
5253 instruction numbers for the links to the previous and next instructions
5256 @item -fdump-translation-unit @r{(C++ only)}
5257 @itemx -fdump-translation-unit-@var{options} @r{(C++ only)}
5258 @opindex fdump-translation-unit
5259 Dump a representation of the tree structure for the entire translation
5260 unit to a file. The file name is made by appending @file{.tu} to the
5261 source file name, and the file is created in the same directory as the
5262 output file. If the @samp{-@var{options}} form is used, @var{options}
5263 controls the details of the dump as described for the
5264 @option{-fdump-tree} options.
5266 @item -fdump-class-hierarchy @r{(C++ only)}
5267 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
5268 @opindex fdump-class-hierarchy
5269 Dump a representation of each class's hierarchy and virtual function
5270 table layout to a file. The file name is made by appending
5271 @file{.class} to the source file name, and the file is created in the
5272 same directory as the output file. If the @samp{-@var{options}} form
5273 is used, @var{options} controls the details of the dump as described
5274 for the @option{-fdump-tree} options.
5276 @item -fdump-ipa-@var{switch}
5278 Control the dumping at various stages of inter-procedural analysis
5279 language tree to a file. The file name is generated by appending a
5280 switch specific suffix to the source file name, and the file is created
5281 in the same directory as the output file. The following dumps are
5286 Enables all inter-procedural analysis dumps.
5289 Dumps information about call-graph optimization, unused function removal,
5290 and inlining decisions.
5293 Dump after function inlining.
5297 @item -fdump-statistics-@var{option}
5298 @opindex fdump-statistics
5299 Enable and control dumping of pass statistics in a separate file. The
5300 file name is generated by appending a suffix ending in
5301 @samp{.statistics} to the source file name, and the file is created in
5302 the same directory as the output file. If the @samp{-@var{option}}
5303 form is used, @samp{-stats} will cause counters to be summed over the
5304 whole compilation unit while @samp{-details} will dump every event as
5305 the passes generate them. The default with no option is to sum
5306 counters for each function compiled.
5308 @item -fdump-tree-@var{switch}
5309 @itemx -fdump-tree-@var{switch}-@var{options}
5311 Control the dumping at various stages of processing the intermediate
5312 language tree to a file. The file name is generated by appending a
5313 switch specific suffix to the source file name, and the file is
5314 created in the same directory as the output file. If the
5315 @samp{-@var{options}} form is used, @var{options} is a list of
5316 @samp{-} separated options that control the details of the dump. Not
5317 all options are applicable to all dumps, those which are not
5318 meaningful will be ignored. The following options are available
5322 Print the address of each node. Usually this is not meaningful as it
5323 changes according to the environment and source file. Its primary use
5324 is for tying up a dump file with a debug environment.
5326 If @code{DECL_ASSEMBLER_NAME} has been set for a given decl, use that
5327 in the dump instead of @code{DECL_NAME}. Its primary use is ease of
5328 use working backward from mangled names in the assembly file.
5330 Inhibit dumping of members of a scope or body of a function merely
5331 because that scope has been reached. Only dump such items when they
5332 are directly reachable by some other path. When dumping pretty-printed
5333 trees, this option inhibits dumping the bodies of control structures.
5335 Print a raw representation of the tree. By default, trees are
5336 pretty-printed into a C-like representation.
5338 Enable more detailed dumps (not honored by every dump option).
5340 Enable dumping various statistics about the pass (not honored by every dump
5343 Enable showing basic block boundaries (disabled in raw dumps).
5345 Enable showing virtual operands for every statement.
5347 Enable showing line numbers for statements.
5349 Enable showing the unique ID (@code{DECL_UID}) for each variable.
5351 Enable showing the tree dump for each statement.
5353 Enable showing the EH region number holding each statement.
5355 Turn on all options, except @option{raw}, @option{slim}, @option{verbose}
5356 and @option{lineno}.
5359 The following tree dumps are possible:
5363 @opindex fdump-tree-original
5364 Dump before any tree based optimization, to @file{@var{file}.original}.
5367 @opindex fdump-tree-optimized
5368 Dump after all tree based optimization, to @file{@var{file}.optimized}.
5371 @opindex fdump-tree-gimple
5372 Dump each function before and after the gimplification pass to a file. The
5373 file name is made by appending @file{.gimple} to the source file name.
5376 @opindex fdump-tree-cfg
5377 Dump the control flow graph of each function to a file. The file name is
5378 made by appending @file{.cfg} to the source file name.
5381 @opindex fdump-tree-vcg
5382 Dump the control flow graph of each function to a file in VCG format. The
5383 file name is made by appending @file{.vcg} to the source file name. Note
5384 that if the file contains more than one function, the generated file cannot
5385 be used directly by VCG@. You will need to cut and paste each function's
5386 graph into its own separate file first.
5389 @opindex fdump-tree-ch
5390 Dump each function after copying loop headers. The file name is made by
5391 appending @file{.ch} to the source file name.
5394 @opindex fdump-tree-ssa
5395 Dump SSA related information to a file. The file name is made by appending
5396 @file{.ssa} to the source file name.
5399 @opindex fdump-tree-alias
5400 Dump aliasing information for each function. The file name is made by
5401 appending @file{.alias} to the source file name.
5404 @opindex fdump-tree-ccp
5405 Dump each function after CCP@. The file name is made by appending
5406 @file{.ccp} to the source file name.
5409 @opindex fdump-tree-storeccp
5410 Dump each function after STORE-CCP@. The file name is made by appending
5411 @file{.storeccp} to the source file name.
5414 @opindex fdump-tree-pre
5415 Dump trees after partial redundancy elimination. The file name is made
5416 by appending @file{.pre} to the source file name.
5419 @opindex fdump-tree-fre
5420 Dump trees after full redundancy elimination. The file name is made
5421 by appending @file{.fre} to the source file name.
5424 @opindex fdump-tree-copyprop
5425 Dump trees after copy propagation. The file name is made
5426 by appending @file{.copyprop} to the source file name.
5428 @item store_copyprop
5429 @opindex fdump-tree-store_copyprop
5430 Dump trees after store copy-propagation. The file name is made
5431 by appending @file{.store_copyprop} to the source file name.
5434 @opindex fdump-tree-dce
5435 Dump each function after dead code elimination. The file name is made by
5436 appending @file{.dce} to the source file name.
5439 @opindex fdump-tree-mudflap
5440 Dump each function after adding mudflap instrumentation. The file name is
5441 made by appending @file{.mudflap} to the source file name.
5444 @opindex fdump-tree-sra
5445 Dump each function after performing scalar replacement of aggregates. The
5446 file name is made by appending @file{.sra} to the source file name.
5449 @opindex fdump-tree-sink
5450 Dump each function after performing code sinking. The file name is made
5451 by appending @file{.sink} to the source file name.
5454 @opindex fdump-tree-dom
5455 Dump each function after applying dominator tree optimizations. The file
5456 name is made by appending @file{.dom} to the source file name.
5459 @opindex fdump-tree-dse
5460 Dump each function after applying dead store elimination. The file
5461 name is made by appending @file{.dse} to the source file name.
5464 @opindex fdump-tree-phiopt
5465 Dump each function after optimizing PHI nodes into straightline code. The file
5466 name is made by appending @file{.phiopt} to the source file name.
5469 @opindex fdump-tree-forwprop
5470 Dump each function after forward propagating single use variables. The file
5471 name is made by appending @file{.forwprop} to the source file name.
5474 @opindex fdump-tree-copyrename
5475 Dump each function after applying the copy rename optimization. The file
5476 name is made by appending @file{.copyrename} to the source file name.
5479 @opindex fdump-tree-nrv
5480 Dump each function after applying the named return value optimization on
5481 generic trees. The file name is made by appending @file{.nrv} to the source
5485 @opindex fdump-tree-vect
5486 Dump each function after applying vectorization of loops. The file name is
5487 made by appending @file{.vect} to the source file name.
5490 @opindex fdump-tree-slp
5491 Dump each function after applying vectorization of basic blocks. The file name
5492 is made by appending @file{.slp} to the source file name.
5495 @opindex fdump-tree-vrp
5496 Dump each function after Value Range Propagation (VRP). The file name
5497 is made by appending @file{.vrp} to the source file name.
5500 @opindex fdump-tree-all
5501 Enable all the available tree dumps with the flags provided in this option.
5504 @item -ftree-vectorizer-verbose=@var{n}
5505 @opindex ftree-vectorizer-verbose
5506 This option controls the amount of debugging output the vectorizer prints.
5507 This information is written to standard error, unless
5508 @option{-fdump-tree-all} or @option{-fdump-tree-vect} is specified,
5509 in which case it is output to the usual dump listing file, @file{.vect}.
5510 For @var{n}=0 no diagnostic information is reported.
5511 If @var{n}=1 the vectorizer reports each loop that got vectorized,
5512 and the total number of loops that got vectorized.
5513 If @var{n}=2 the vectorizer also reports non-vectorized loops that passed
5514 the first analysis phase (vect_analyze_loop_form) - i.e.@: countable,
5515 inner-most, single-bb, single-entry/exit loops. This is the same verbosity
5516 level that @option{-fdump-tree-vect-stats} uses.
5517 Higher verbosity levels mean either more information dumped for each
5518 reported loop, or same amount of information reported for more loops:
5519 if @var{n}=3, vectorizer cost model information is reported.
5520 If @var{n}=4, alignment related information is added to the reports.
5521 If @var{n}=5, data-references related information (e.g.@: memory dependences,
5522 memory access-patterns) is added to the reports.
5523 If @var{n}=6, the vectorizer reports also non-vectorized inner-most loops
5524 that did not pass the first analysis phase (i.e., may not be countable, or
5525 may have complicated control-flow).
5526 If @var{n}=7, the vectorizer reports also non-vectorized nested loops.
5527 If @var{n}=8, SLP related information is added to the reports.
5528 For @var{n}=9, all the information the vectorizer generates during its
5529 analysis and transformation is reported. This is the same verbosity level
5530 that @option{-fdump-tree-vect-details} uses.
5532 @item -frandom-seed=@var{string}
5533 @opindex frandom-seed
5534 This option provides a seed that GCC uses when it would otherwise use
5535 random numbers. It is used to generate certain symbol names
5536 that have to be different in every compiled file. It is also used to
5537 place unique stamps in coverage data files and the object files that
5538 produce them. You can use the @option{-frandom-seed} option to produce
5539 reproducibly identical object files.
5541 The @var{string} should be different for every file you compile.
5543 @item -fsched-verbose=@var{n}
5544 @opindex fsched-verbose
5545 On targets that use instruction scheduling, this option controls the
5546 amount of debugging output the scheduler prints. This information is
5547 written to standard error, unless @option{-fdump-rtl-sched1} or
5548 @option{-fdump-rtl-sched2} is specified, in which case it is output
5549 to the usual dump listing file, @file{.sched1} or @file{.sched2}
5550 respectively. However for @var{n} greater than nine, the output is
5551 always printed to standard error.
5553 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
5554 same information as @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2}.
5555 For @var{n} greater than one, it also output basic block probabilities,
5556 detailed ready list information and unit/insn info. For @var{n} greater
5557 than two, it includes RTL at abort point, control-flow and regions info.
5558 And for @var{n} over four, @option{-fsched-verbose} also includes
5562 @itemx -save-temps=cwd
5564 Store the usual ``temporary'' intermediate files permanently; place them
5565 in the current directory and name them based on the source file. Thus,
5566 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
5567 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
5568 preprocessed @file{foo.i} output file even though the compiler now
5569 normally uses an integrated preprocessor.
5571 When used in combination with the @option{-x} command line option,
5572 @option{-save-temps} is sensible enough to avoid over writing an
5573 input source file with the same extension as an intermediate file.
5574 The corresponding intermediate file may be obtained by renaming the
5575 source file before using @option{-save-temps}.
5577 If you invoke GCC in parallel, compiling several different source
5578 files that share a common base name in different subdirectories or the
5579 same source file compiled for multiple output destinations, it is
5580 likely that the different parallel compilers will interfere with each
5581 other, and overwrite the temporary files. For instance:
5584 gcc -save-temps -o outdir1/foo.o indir1/foo.c&
5585 gcc -save-temps -o outdir2/foo.o indir2/foo.c&
5588 may result in @file{foo.i} and @file{foo.o} being written to
5589 simultaneously by both compilers.
5591 @item -save-temps=obj
5592 @opindex save-temps=obj
5593 Store the usual ``temporary'' intermediate files permanently. If the
5594 @option{-o} option is used, the temporary files are based on the
5595 object file. If the @option{-o} option is not used, the
5596 @option{-save-temps=obj} switch behaves like @option{-save-temps}.
5601 gcc -save-temps=obj -c foo.c
5602 gcc -save-temps=obj -c bar.c -o dir/xbar.o
5603 gcc -save-temps=obj foobar.c -o dir2/yfoobar
5606 would create @file{foo.i}, @file{foo.s}, @file{dir/xbar.i},
5607 @file{dir/xbar.s}, @file{dir2/yfoobar.i}, @file{dir2/yfoobar.s}, and
5608 @file{dir2/yfoobar.o}.
5610 @item -time@r{[}=@var{file}@r{]}
5612 Report the CPU time taken by each subprocess in the compilation
5613 sequence. For C source files, this is the compiler proper and assembler
5614 (plus the linker if linking is done).
5616 Without the specification of an output file, the output looks like this:
5623 The first number on each line is the ``user time'', that is time spent
5624 executing the program itself. The second number is ``system time'',
5625 time spent executing operating system routines on behalf of the program.
5626 Both numbers are in seconds.
5628 With the specification of an output file, the output is appended to the
5629 named file, and it looks like this:
5632 0.12 0.01 cc1 @var{options}
5633 0.00 0.01 as @var{options}
5636 The ``user time'' and the ``system time'' are moved before the program
5637 name, and the options passed to the program are displayed, so that one
5638 can later tell what file was being compiled, and with which options.
5640 @item -fvar-tracking
5641 @opindex fvar-tracking
5642 Run variable tracking pass. It computes where variables are stored at each
5643 position in code. Better debugging information is then generated
5644 (if the debugging information format supports this information).
5646 It is enabled by default when compiling with optimization (@option{-Os},
5647 @option{-O}, @option{-O2}, @dots{}), debugging information (@option{-g}) and
5648 the debug info format supports it.
5650 @item -fvar-tracking-assignments
5651 @opindex fvar-tracking-assignments
5652 @opindex fno-var-tracking-assignments
5653 Annotate assignments to user variables early in the compilation and
5654 attempt to carry the annotations over throughout the compilation all the
5655 way to the end, in an attempt to improve debug information while
5656 optimizing. Use of @option{-gdwarf-4} is recommended along with it.
5658 It can be enabled even if var-tracking is disabled, in which case
5659 annotations will be created and maintained, but discarded at the end.
5661 @item -fvar-tracking-assignments-toggle
5662 @opindex fvar-tracking-assignments-toggle
5663 @opindex fno-var-tracking-assignments-toggle
5664 Toggle @option{-fvar-tracking-assignments}, in the same way that
5665 @option{-gtoggle} toggles @option{-g}.
5667 @item -print-file-name=@var{library}
5668 @opindex print-file-name
5669 Print the full absolute name of the library file @var{library} that
5670 would be used when linking---and don't do anything else. With this
5671 option, GCC does not compile or link anything; it just prints the
5674 @item -print-multi-directory
5675 @opindex print-multi-directory
5676 Print the directory name corresponding to the multilib selected by any
5677 other switches present in the command line. This directory is supposed
5678 to exist in @env{GCC_EXEC_PREFIX}.
5680 @item -print-multi-lib
5681 @opindex print-multi-lib
5682 Print the mapping from multilib directory names to compiler switches
5683 that enable them. The directory name is separated from the switches by
5684 @samp{;}, and each switch starts with an @samp{@@} instead of the
5685 @samp{-}, without spaces between multiple switches. This is supposed to
5686 ease shell-processing.
5688 @item -print-multi-os-directory
5689 @opindex print-multi-os-directory
5690 Print the path to OS libraries for the selected
5691 multilib, relative to some @file{lib} subdirectory. If OS libraries are
5692 present in the @file{lib} subdirectory and no multilibs are used, this is
5693 usually just @file{.}, if OS libraries are present in @file{lib@var{suffix}}
5694 sibling directories this prints e.g.@: @file{../lib64}, @file{../lib} or
5695 @file{../lib32}, or if OS libraries are present in @file{lib/@var{subdir}}
5696 subdirectories it prints e.g.@: @file{amd64}, @file{sparcv9} or @file{ev6}.
5698 @item -print-prog-name=@var{program}
5699 @opindex print-prog-name
5700 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
5702 @item -print-libgcc-file-name
5703 @opindex print-libgcc-file-name
5704 Same as @option{-print-file-name=libgcc.a}.
5706 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
5707 but you do want to link with @file{libgcc.a}. You can do
5710 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
5713 @item -print-search-dirs
5714 @opindex print-search-dirs
5715 Print the name of the configured installation directory and a list of
5716 program and library directories @command{gcc} will search---and don't do anything else.
5718 This is useful when @command{gcc} prints the error message
5719 @samp{installation problem, cannot exec cpp0: No such file or directory}.
5720 To resolve this you either need to put @file{cpp0} and the other compiler
5721 components where @command{gcc} expects to find them, or you can set the environment
5722 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
5723 Don't forget the trailing @samp{/}.
5724 @xref{Environment Variables}.
5726 @item -print-sysroot
5727 @opindex print-sysroot
5728 Print the target sysroot directory that will be used during
5729 compilation. This is the target sysroot specified either at configure
5730 time or using the @option{--sysroot} option, possibly with an extra
5731 suffix that depends on compilation options. If no target sysroot is
5732 specified, the option prints nothing.
5734 @item -print-sysroot-headers-suffix
5735 @opindex print-sysroot-headers-suffix
5736 Print the suffix added to the target sysroot when searching for
5737 headers, or give an error if the compiler is not configured with such
5738 a suffix---and don't do anything else.
5741 @opindex dumpmachine
5742 Print the compiler's target machine (for example,
5743 @samp{i686-pc-linux-gnu})---and don't do anything else.
5746 @opindex dumpversion
5747 Print the compiler version (for example, @samp{3.0})---and don't do
5752 Print the compiler's built-in specs---and don't do anything else. (This
5753 is used when GCC itself is being built.) @xref{Spec Files}.
5755 @item -feliminate-unused-debug-types
5756 @opindex feliminate-unused-debug-types
5757 Normally, when producing DWARF2 output, GCC will emit debugging
5758 information for all types declared in a compilation
5759 unit, regardless of whether or not they are actually used
5760 in that compilation unit. Sometimes this is useful, such as
5761 if, in the debugger, you want to cast a value to a type that is
5762 not actually used in your program (but is declared). More often,
5763 however, this results in a significant amount of wasted space.
5764 With this option, GCC will avoid producing debug symbol output
5765 for types that are nowhere used in the source file being compiled.
5768 @node Optimize Options
5769 @section Options That Control Optimization
5770 @cindex optimize options
5771 @cindex options, optimization
5773 These options control various sorts of optimizations.
5775 Without any optimization option, the compiler's goal is to reduce the
5776 cost of compilation and to make debugging produce the expected
5777 results. Statements are independent: if you stop the program with a
5778 breakpoint between statements, you can then assign a new value to any
5779 variable or change the program counter to any other statement in the
5780 function and get exactly the results you would expect from the source
5783 Turning on optimization flags makes the compiler attempt to improve
5784 the performance and/or code size at the expense of compilation time
5785 and possibly the ability to debug the program.
5787 The compiler performs optimization based on the knowledge it has of the
5788 program. Compiling multiple files at once to a single output file mode allows
5789 the compiler to use information gained from all of the files when compiling
5792 Not all optimizations are controlled directly by a flag. Only
5793 optimizations that have a flag are listed in this section.
5795 Most optimizations are only enabled if an @option{-O} level is set on
5796 the command line. Otherwise they are disabled, even if individual
5797 optimization flags are specified.
5799 Depending on the target and how GCC was configured, a slightly different
5800 set of optimizations may be enabled at each @option{-O} level than
5801 those listed here. You can invoke GCC with @samp{-Q --help=optimizers}
5802 to find out the exact set of optimizations that are enabled at each level.
5803 @xref{Overall Options}, for examples.
5810 Optimize. Optimizing compilation takes somewhat more time, and a lot
5811 more memory for a large function.
5813 With @option{-O}, the compiler tries to reduce code size and execution
5814 time, without performing any optimizations that take a great deal of
5817 @option{-O} turns on the following optimization flags:
5820 -fcprop-registers @gol
5823 -fdelayed-branch @gol
5825 -fguess-branch-probability @gol
5826 -fif-conversion2 @gol
5827 -fif-conversion @gol
5828 -fipa-pure-const @gol
5830 -fipa-reference @gol
5832 -fsplit-wide-types @gol
5833 -ftree-builtin-call-dce @gol
5836 -ftree-copyrename @gol
5838 -ftree-dominator-opts @gol
5840 -ftree-forwprop @gol
5848 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
5849 where doing so does not interfere with debugging.
5853 Optimize even more. GCC performs nearly all supported optimizations
5854 that do not involve a space-speed tradeoff.
5855 As compared to @option{-O}, this option increases both compilation time
5856 and the performance of the generated code.
5858 @option{-O2} turns on all optimization flags specified by @option{-O}. It
5859 also turns on the following optimization flags:
5860 @gccoptlist{-fthread-jumps @gol
5861 -falign-functions -falign-jumps @gol
5862 -falign-loops -falign-labels @gol
5865 -fcse-follow-jumps -fcse-skip-blocks @gol
5866 -fdelete-null-pointer-checks @gol
5867 -fexpensive-optimizations @gol
5868 -fgcse -fgcse-lm @gol
5869 -finline-small-functions @gol
5870 -findirect-inlining @gol
5872 -foptimize-sibling-calls @gol
5875 -freorder-blocks -freorder-functions @gol
5876 -frerun-cse-after-loop @gol
5877 -fsched-interblock -fsched-spec @gol
5878 -fschedule-insns -fschedule-insns2 @gol
5879 -fstrict-aliasing -fstrict-overflow @gol
5880 -ftree-switch-conversion @gol
5884 Please note the warning under @option{-fgcse} about
5885 invoking @option{-O2} on programs that use computed gotos.
5889 Optimize yet more. @option{-O3} turns on all optimizations specified
5890 by @option{-O2} and also turns on the @option{-finline-functions},
5891 @option{-funswitch-loops}, @option{-fpredictive-commoning},
5892 @option{-fgcse-after-reload} and @option{-ftree-vectorize} options.
5896 Reduce compilation time and make debugging produce the expected
5897 results. This is the default.
5901 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
5902 do not typically increase code size. It also performs further
5903 optimizations designed to reduce code size.
5905 @option{-Os} disables the following optimization flags:
5906 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
5907 -falign-labels -freorder-blocks -freorder-blocks-and-partition @gol
5908 -fprefetch-loop-arrays -ftree-vect-loop-version}
5912 Disregard strict standards compliance. @option{-Ofast} enables all
5913 @option{-O3} optimizations. It also enables optimizations that are not
5914 valid for all standard compliant programs.
5915 It turns on @option{-ffast-math}.
5917 If you use multiple @option{-O} options, with or without level numbers,
5918 the last such option is the one that is effective.
5921 Options of the form @option{-f@var{flag}} specify machine-independent
5922 flags. Most flags have both positive and negative forms; the negative
5923 form of @option{-ffoo} would be @option{-fno-foo}. In the table
5924 below, only one of the forms is listed---the one you typically will
5925 use. You can figure out the other form by either removing @samp{no-}
5928 The following options control specific optimizations. They are either
5929 activated by @option{-O} options or are related to ones that are. You
5930 can use the following flags in the rare cases when ``fine-tuning'' of
5931 optimizations to be performed is desired.
5934 @item -fno-default-inline
5935 @opindex fno-default-inline
5936 Do not make member functions inline by default merely because they are
5937 defined inside the class scope (C++ only). Otherwise, when you specify
5938 @w{@option{-O}}, member functions defined inside class scope are compiled
5939 inline by default; i.e., you don't need to add @samp{inline} in front of
5940 the member function name.
5942 @item -fno-defer-pop
5943 @opindex fno-defer-pop
5944 Always pop the arguments to each function call as soon as that function
5945 returns. For machines which must pop arguments after a function call,
5946 the compiler normally lets arguments accumulate on the stack for several
5947 function calls and pops them all at once.
5949 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5951 @item -fforward-propagate
5952 @opindex fforward-propagate
5953 Perform a forward propagation pass on RTL@. The pass tries to combine two
5954 instructions and checks if the result can be simplified. If loop unrolling
5955 is active, two passes are performed and the second is scheduled after
5958 This option is enabled by default at optimization levels @option{-O},
5959 @option{-O2}, @option{-O3}, @option{-Os}.
5961 @item -fomit-frame-pointer
5962 @opindex fomit-frame-pointer
5963 Don't keep the frame pointer in a register for functions that
5964 don't need one. This avoids the instructions to save, set up and
5965 restore frame pointers; it also makes an extra register available
5966 in many functions. @strong{It also makes debugging impossible on
5969 On some machines, such as the VAX, this flag has no effect, because
5970 the standard calling sequence automatically handles the frame pointer
5971 and nothing is saved by pretending it doesn't exist. The
5972 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
5973 whether a target machine supports this flag. @xref{Registers,,Register
5974 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
5976 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5978 @item -foptimize-sibling-calls
5979 @opindex foptimize-sibling-calls
5980 Optimize sibling and tail recursive calls.
5982 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5986 Don't pay attention to the @code{inline} keyword. Normally this option
5987 is used to keep the compiler from expanding any functions inline.
5988 Note that if you are not optimizing, no functions can be expanded inline.
5990 @item -finline-small-functions
5991 @opindex finline-small-functions
5992 Integrate functions into their callers when their body is smaller than expected
5993 function call code (so overall size of program gets smaller). The compiler
5994 heuristically decides which functions are simple enough to be worth integrating
5997 Enabled at level @option{-O2}.
5999 @item -findirect-inlining
6000 @opindex findirect-inlining
6001 Inline also indirect calls that are discovered to be known at compile
6002 time thanks to previous inlining. This option has any effect only
6003 when inlining itself is turned on by the @option{-finline-functions}
6004 or @option{-finline-small-functions} options.
6006 Enabled at level @option{-O2}.
6008 @item -finline-functions
6009 @opindex finline-functions
6010 Integrate all simple functions into their callers. The compiler
6011 heuristically decides which functions are simple enough to be worth
6012 integrating in this way.
6014 If all calls to a given function are integrated, and the function is
6015 declared @code{static}, then the function is normally not output as
6016 assembler code in its own right.
6018 Enabled at level @option{-O3}.
6020 @item -finline-functions-called-once
6021 @opindex finline-functions-called-once
6022 Consider all @code{static} functions called once for inlining into their
6023 caller even if they are not marked @code{inline}. If a call to a given
6024 function is integrated, then the function is not output as assembler code
6027 Enabled at levels @option{-O1}, @option{-O2}, @option{-O3} and @option{-Os}.
6029 @item -fearly-inlining
6030 @opindex fearly-inlining
6031 Inline functions marked by @code{always_inline} and functions whose body seems
6032 smaller than the function call overhead early before doing
6033 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
6034 makes profiling significantly cheaper and usually inlining faster on programs
6035 having large chains of nested wrapper functions.
6041 Perform interprocedural scalar replacement of aggregates, removal of
6042 unused parameters and replacement of parameters passed by reference
6043 by parameters passed by value.
6045 Enabled at levels @option{-O2}, @option{-O3} and @option{-Os}.
6047 @item -finline-limit=@var{n}
6048 @opindex finline-limit
6049 By default, GCC limits the size of functions that can be inlined. This flag
6050 allows coarse control of this limit. @var{n} is the size of functions that
6051 can be inlined in number of pseudo instructions.
6053 Inlining is actually controlled by a number of parameters, which may be
6054 specified individually by using @option{--param @var{name}=@var{value}}.
6055 The @option{-finline-limit=@var{n}} option sets some of these parameters
6059 @item max-inline-insns-single
6060 is set to @var{n}/2.
6061 @item max-inline-insns-auto
6062 is set to @var{n}/2.
6065 See below for a documentation of the individual
6066 parameters controlling inlining and for the defaults of these parameters.
6068 @emph{Note:} there may be no value to @option{-finline-limit} that results
6069 in default behavior.
6071 @emph{Note:} pseudo instruction represents, in this particular context, an
6072 abstract measurement of function's size. In no way does it represent a count
6073 of assembly instructions and as such its exact meaning might change from one
6074 release to an another.
6076 @item -fkeep-inline-functions
6077 @opindex fkeep-inline-functions
6078 In C, emit @code{static} functions that are declared @code{inline}
6079 into the object file, even if the function has been inlined into all
6080 of its callers. This switch does not affect functions using the
6081 @code{extern inline} extension in GNU C90@. In C++, emit any and all
6082 inline functions into the object file.
6084 @item -fkeep-static-consts
6085 @opindex fkeep-static-consts
6086 Emit variables declared @code{static const} when optimization isn't turned
6087 on, even if the variables aren't referenced.
6089 GCC enables this option by default. If you want to force the compiler to
6090 check if the variable was referenced, regardless of whether or not
6091 optimization is turned on, use the @option{-fno-keep-static-consts} option.
6093 @item -fmerge-constants
6094 @opindex fmerge-constants
6095 Attempt to merge identical constants (string constants and floating point
6096 constants) across compilation units.
6098 This option is the default for optimized compilation if the assembler and
6099 linker support it. Use @option{-fno-merge-constants} to inhibit this
6102 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6104 @item -fmerge-all-constants
6105 @opindex fmerge-all-constants
6106 Attempt to merge identical constants and identical variables.
6108 This option implies @option{-fmerge-constants}. In addition to
6109 @option{-fmerge-constants} this considers e.g.@: even constant initialized
6110 arrays or initialized constant variables with integral or floating point
6111 types. Languages like C or C++ require each variable, including multiple
6112 instances of the same variable in recursive calls, to have distinct locations,
6113 so using this option will result in non-conforming
6116 @item -fmodulo-sched
6117 @opindex fmodulo-sched
6118 Perform swing modulo scheduling immediately before the first scheduling
6119 pass. This pass looks at innermost loops and reorders their
6120 instructions by overlapping different iterations.
6122 @item -fmodulo-sched-allow-regmoves
6123 @opindex fmodulo-sched-allow-regmoves
6124 Perform more aggressive SMS based modulo scheduling with register moves
6125 allowed. By setting this flag certain anti-dependences edges will be
6126 deleted which will trigger the generation of reg-moves based on the
6127 life-range analysis. This option is effective only with
6128 @option{-fmodulo-sched} enabled.
6130 @item -fno-branch-count-reg
6131 @opindex fno-branch-count-reg
6132 Do not use ``decrement and branch'' instructions on a count register,
6133 but instead generate a sequence of instructions that decrement a
6134 register, compare it against zero, then branch based upon the result.
6135 This option is only meaningful on architectures that support such
6136 instructions, which include x86, PowerPC, IA-64 and S/390.
6138 The default is @option{-fbranch-count-reg}.
6140 @item -fno-function-cse
6141 @opindex fno-function-cse
6142 Do not put function addresses in registers; make each instruction that
6143 calls a constant function contain the function's address explicitly.
6145 This option results in less efficient code, but some strange hacks
6146 that alter the assembler output may be confused by the optimizations
6147 performed when this option is not used.
6149 The default is @option{-ffunction-cse}
6151 @item -fno-zero-initialized-in-bss
6152 @opindex fno-zero-initialized-in-bss
6153 If the target supports a BSS section, GCC by default puts variables that
6154 are initialized to zero into BSS@. This can save space in the resulting
6157 This option turns off this behavior because some programs explicitly
6158 rely on variables going to the data section. E.g., so that the
6159 resulting executable can find the beginning of that section and/or make
6160 assumptions based on that.
6162 The default is @option{-fzero-initialized-in-bss}.
6164 @item -fmudflap -fmudflapth -fmudflapir
6168 @cindex bounds checking
6170 For front-ends that support it (C and C++), instrument all risky
6171 pointer/array dereferencing operations, some standard library
6172 string/heap functions, and some other associated constructs with
6173 range/validity tests. Modules so instrumented should be immune to
6174 buffer overflows, invalid heap use, and some other classes of C/C++
6175 programming errors. The instrumentation relies on a separate runtime
6176 library (@file{libmudflap}), which will be linked into a program if
6177 @option{-fmudflap} is given at link time. Run-time behavior of the
6178 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
6179 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
6182 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
6183 link if your program is multi-threaded. Use @option{-fmudflapir}, in
6184 addition to @option{-fmudflap} or @option{-fmudflapth}, if
6185 instrumentation should ignore pointer reads. This produces less
6186 instrumentation (and therefore faster execution) and still provides
6187 some protection against outright memory corrupting writes, but allows
6188 erroneously read data to propagate within a program.
6190 @item -fthread-jumps
6191 @opindex fthread-jumps
6192 Perform optimizations where we check to see if a jump branches to a
6193 location where another comparison subsumed by the first is found. If
6194 so, the first branch is redirected to either the destination of the
6195 second branch or a point immediately following it, depending on whether
6196 the condition is known to be true or false.
6198 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6200 @item -fsplit-wide-types
6201 @opindex fsplit-wide-types
6202 When using a type that occupies multiple registers, such as @code{long
6203 long} on a 32-bit system, split the registers apart and allocate them
6204 independently. This normally generates better code for those types,
6205 but may make debugging more difficult.
6207 Enabled at levels @option{-O}, @option{-O2}, @option{-O3},
6210 @item -fcse-follow-jumps
6211 @opindex fcse-follow-jumps
6212 In common subexpression elimination (CSE), scan through jump instructions
6213 when the target of the jump is not reached by any other path. For
6214 example, when CSE encounters an @code{if} statement with an
6215 @code{else} clause, CSE will follow the jump when the condition
6218 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6220 @item -fcse-skip-blocks
6221 @opindex fcse-skip-blocks
6222 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
6223 follow jumps which conditionally skip over blocks. When CSE
6224 encounters a simple @code{if} statement with no else clause,
6225 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
6226 body of the @code{if}.
6228 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6230 @item -frerun-cse-after-loop
6231 @opindex frerun-cse-after-loop
6232 Re-run common subexpression elimination after loop optimizations has been
6235 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6239 Perform a global common subexpression elimination pass.
6240 This pass also performs global constant and copy propagation.
6242 @emph{Note:} When compiling a program using computed gotos, a GCC
6243 extension, you may get better runtime performance if you disable
6244 the global common subexpression elimination pass by adding
6245 @option{-fno-gcse} to the command line.
6247 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6251 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
6252 attempt to move loads which are only killed by stores into themselves. This
6253 allows a loop containing a load/store sequence to be changed to a load outside
6254 the loop, and a copy/store within the loop.
6256 Enabled by default when gcse is enabled.
6260 When @option{-fgcse-sm} is enabled, a store motion pass is run after
6261 global common subexpression elimination. This pass will attempt to move
6262 stores out of loops. When used in conjunction with @option{-fgcse-lm},
6263 loops containing a load/store sequence can be changed to a load before
6264 the loop and a store after the loop.
6266 Not enabled at any optimization level.
6270 When @option{-fgcse-las} is enabled, the global common subexpression
6271 elimination pass eliminates redundant loads that come after stores to the
6272 same memory location (both partial and full redundancies).
6274 Not enabled at any optimization level.
6276 @item -fgcse-after-reload
6277 @opindex fgcse-after-reload
6278 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
6279 pass is performed after reload. The purpose of this pass is to cleanup
6282 @item -funsafe-loop-optimizations
6283 @opindex funsafe-loop-optimizations
6284 If given, the loop optimizer will assume that loop indices do not
6285 overflow, and that the loops with nontrivial exit condition are not
6286 infinite. This enables a wider range of loop optimizations even if
6287 the loop optimizer itself cannot prove that these assumptions are valid.
6288 Using @option{-Wunsafe-loop-optimizations}, the compiler will warn you
6289 if it finds this kind of loop.
6291 @item -fcrossjumping
6292 @opindex fcrossjumping
6293 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
6294 resulting code may or may not perform better than without cross-jumping.
6296 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6298 @item -fauto-inc-dec
6299 @opindex fauto-inc-dec
6300 Combine increments or decrements of addresses with memory accesses.
6301 This pass is always skipped on architectures that do not have
6302 instructions to support this. Enabled by default at @option{-O} and
6303 higher on architectures that support this.
6307 Perform dead code elimination (DCE) on RTL@.
6308 Enabled by default at @option{-O} and higher.
6312 Perform dead store elimination (DSE) on RTL@.
6313 Enabled by default at @option{-O} and higher.
6315 @item -fif-conversion
6316 @opindex fif-conversion
6317 Attempt to transform conditional jumps into branch-less equivalents. This
6318 include use of conditional moves, min, max, set flags and abs instructions, and
6319 some tricks doable by standard arithmetics. The use of conditional execution
6320 on chips where it is available is controlled by @code{if-conversion2}.
6322 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6324 @item -fif-conversion2
6325 @opindex fif-conversion2
6326 Use conditional execution (where available) to transform conditional jumps into
6327 branch-less equivalents.
6329 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6331 @item -fdelete-null-pointer-checks
6332 @opindex fdelete-null-pointer-checks
6333 Assume that programs cannot safely dereference null pointers, and that
6334 no code or data element resides there. This enables simple constant
6335 folding optimizations at all optimization levels. In addition, other
6336 optimization passes in GCC use this flag to control global dataflow
6337 analyses that eliminate useless checks for null pointers; these assume
6338 that if a pointer is checked after it has already been dereferenced,
6341 Note however that in some environments this assumption is not true.
6342 Use @option{-fno-delete-null-pointer-checks} to disable this optimization
6343 for programs which depend on that behavior.
6345 Some targets, especially embedded ones, disable this option at all levels.
6346 Otherwise it is enabled at all levels: @option{-O0}, @option{-O1},
6347 @option{-O2}, @option{-O3}, @option{-Os}. Passes that use the information
6348 are enabled independently at different optimization levels.
6350 @item -fexpensive-optimizations
6351 @opindex fexpensive-optimizations
6352 Perform a number of minor optimizations that are relatively expensive.
6354 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6356 @item -foptimize-register-move
6358 @opindex foptimize-register-move
6360 Attempt to reassign register numbers in move instructions and as
6361 operands of other simple instructions in order to maximize the amount of
6362 register tying. This is especially helpful on machines with two-operand
6365 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
6368 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6370 @item -fira-algorithm=@var{algorithm}
6371 Use specified coloring algorithm for the integrated register
6372 allocator. The @var{algorithm} argument should be @code{priority} or
6373 @code{CB}. The first algorithm specifies Chow's priority coloring,
6374 the second one specifies Chaitin-Briggs coloring. The second
6375 algorithm can be unimplemented for some architectures. If it is
6376 implemented, it is the default because Chaitin-Briggs coloring as a
6377 rule generates a better code.
6379 @item -fira-region=@var{region}
6380 Use specified regions for the integrated register allocator. The
6381 @var{region} argument should be one of @code{all}, @code{mixed}, or
6382 @code{one}. The first value means using all loops as register
6383 allocation regions, the second value which is the default means using
6384 all loops except for loops with small register pressure as the
6385 regions, and third one means using all function as a single region.
6386 The first value can give best result for machines with small size and
6387 irregular register set, the third one results in faster and generates
6388 decent code and the smallest size code, and the default value usually
6389 give the best results in most cases and for most architectures.
6391 @item -fira-coalesce
6392 @opindex fira-coalesce
6393 Do optimistic register coalescing. This option might be profitable for
6394 architectures with big regular register files.
6396 @item -fira-loop-pressure
6397 @opindex fira-loop-pressure
6398 Use IRA to evaluate register pressure in loops for decision to move
6399 loop invariants. Usage of this option usually results in generation
6400 of faster and smaller code on machines with big register files (>= 32
6401 registers) but it can slow compiler down.
6403 This option is enabled at level @option{-O3} for some targets.
6405 @item -fno-ira-share-save-slots
6406 @opindex fno-ira-share-save-slots
6407 Switch off sharing stack slots used for saving call used hard
6408 registers living through a call. Each hard register will get a
6409 separate stack slot and as a result function stack frame will be
6412 @item -fno-ira-share-spill-slots
6413 @opindex fno-ira-share-spill-slots
6414 Switch off sharing stack slots allocated for pseudo-registers. Each
6415 pseudo-register which did not get a hard register will get a separate
6416 stack slot and as a result function stack frame will be bigger.
6418 @item -fira-verbose=@var{n}
6419 @opindex fira-verbose
6420 Set up how verbose dump file for the integrated register allocator
6421 will be. Default value is 5. If the value is greater or equal to 10,
6422 the dump file will be stderr as if the value were @var{n} minus 10.
6424 @item -fdelayed-branch
6425 @opindex fdelayed-branch
6426 If supported for the target machine, attempt to reorder instructions
6427 to exploit instruction slots available after delayed branch
6430 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6432 @item -fschedule-insns
6433 @opindex fschedule-insns
6434 If supported for the target machine, attempt to reorder instructions to
6435 eliminate execution stalls due to required data being unavailable. This
6436 helps machines that have slow floating point or memory load instructions
6437 by allowing other instructions to be issued until the result of the load
6438 or floating point instruction is required.
6440 Enabled at levels @option{-O2}, @option{-O3}.
6442 @item -fschedule-insns2
6443 @opindex fschedule-insns2
6444 Similar to @option{-fschedule-insns}, but requests an additional pass of
6445 instruction scheduling after register allocation has been done. This is
6446 especially useful on machines with a relatively small number of
6447 registers and where memory load instructions take more than one cycle.
6449 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6451 @item -fno-sched-interblock
6452 @opindex fno-sched-interblock
6453 Don't schedule instructions across basic blocks. This is normally
6454 enabled by default when scheduling before register allocation, i.e.@:
6455 with @option{-fschedule-insns} or at @option{-O2} or higher.
6457 @item -fno-sched-spec
6458 @opindex fno-sched-spec
6459 Don't allow speculative motion of non-load instructions. This is normally
6460 enabled by default when scheduling before register allocation, i.e.@:
6461 with @option{-fschedule-insns} or at @option{-O2} or higher.
6463 @item -fsched-pressure
6464 @opindex fsched-pressure
6465 Enable register pressure sensitive insn scheduling before the register
6466 allocation. This only makes sense when scheduling before register
6467 allocation is enabled, i.e.@: with @option{-fschedule-insns} or at
6468 @option{-O2} or higher. Usage of this option can improve the
6469 generated code and decrease its size by preventing register pressure
6470 increase above the number of available hard registers and as a
6471 consequence register spills in the register allocation.
6473 @item -fsched-spec-load
6474 @opindex fsched-spec-load
6475 Allow speculative motion of some load instructions. This only makes
6476 sense when scheduling before register allocation, i.e.@: with
6477 @option{-fschedule-insns} or at @option{-O2} or higher.
6479 @item -fsched-spec-load-dangerous
6480 @opindex fsched-spec-load-dangerous
6481 Allow speculative motion of more load instructions. This only makes
6482 sense when scheduling before register allocation, i.e.@: with
6483 @option{-fschedule-insns} or at @option{-O2} or higher.
6485 @item -fsched-stalled-insns
6486 @itemx -fsched-stalled-insns=@var{n}
6487 @opindex fsched-stalled-insns
6488 Define how many insns (if any) can be moved prematurely from the queue
6489 of stalled insns into the ready list, during the second scheduling pass.
6490 @option{-fno-sched-stalled-insns} means that no insns will be moved
6491 prematurely, @option{-fsched-stalled-insns=0} means there is no limit
6492 on how many queued insns can be moved prematurely.
6493 @option{-fsched-stalled-insns} without a value is equivalent to
6494 @option{-fsched-stalled-insns=1}.
6496 @item -fsched-stalled-insns-dep
6497 @itemx -fsched-stalled-insns-dep=@var{n}
6498 @opindex fsched-stalled-insns-dep
6499 Define how many insn groups (cycles) will be examined for a dependency
6500 on a stalled insn that is candidate for premature removal from the queue
6501 of stalled insns. This has an effect only during the second scheduling pass,
6502 and only if @option{-fsched-stalled-insns} is used.
6503 @option{-fno-sched-stalled-insns-dep} is equivalent to
6504 @option{-fsched-stalled-insns-dep=0}.
6505 @option{-fsched-stalled-insns-dep} without a value is equivalent to
6506 @option{-fsched-stalled-insns-dep=1}.
6508 @item -fsched2-use-superblocks
6509 @opindex fsched2-use-superblocks
6510 When scheduling after register allocation, do use superblock scheduling
6511 algorithm. Superblock scheduling allows motion across basic block boundaries
6512 resulting on faster schedules. This option is experimental, as not all machine
6513 descriptions used by GCC model the CPU closely enough to avoid unreliable
6514 results from the algorithm.
6516 This only makes sense when scheduling after register allocation, i.e.@: with
6517 @option{-fschedule-insns2} or at @option{-O2} or higher.
6519 @item -fsched-group-heuristic
6520 @opindex fsched-group-heuristic
6521 Enable the group heuristic in the scheduler. This heuristic favors
6522 the instruction that belongs to a schedule group. This is enabled
6523 by default when scheduling is enabled, i.e.@: with @option{-fschedule-insns}
6524 or @option{-fschedule-insns2} or at @option{-O2} or higher.
6526 @item -fsched-critical-path-heuristic
6527 @opindex fsched-critical-path-heuristic
6528 Enable the critical-path heuristic in the scheduler. This heuristic favors
6529 instructions on the critical path. This is enabled by default when
6530 scheduling is enabled, i.e.@: with @option{-fschedule-insns}
6531 or @option{-fschedule-insns2} or at @option{-O2} or higher.
6533 @item -fsched-spec-insn-heuristic
6534 @opindex fsched-spec-insn-heuristic
6535 Enable the speculative instruction heuristic in the scheduler. This
6536 heuristic favors speculative instructions with greater dependency weakness.
6537 This is enabled by default when scheduling is enabled, i.e.@:
6538 with @option{-fschedule-insns} or @option{-fschedule-insns2}
6539 or at @option{-O2} or higher.
6541 @item -fsched-rank-heuristic
6542 @opindex fsched-rank-heuristic
6543 Enable the rank heuristic in the scheduler. This heuristic favors
6544 the instruction belonging to a basic block with greater size or frequency.
6545 This is enabled by default when scheduling is enabled, i.e.@:
6546 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6547 at @option{-O2} or higher.
6549 @item -fsched-last-insn-heuristic
6550 @opindex fsched-last-insn-heuristic
6551 Enable the last-instruction heuristic in the scheduler. This heuristic
6552 favors the instruction that is less dependent on the last instruction
6553 scheduled. This is enabled by default when scheduling is enabled,
6554 i.e.@: with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6555 at @option{-O2} or higher.
6557 @item -fsched-dep-count-heuristic
6558 @opindex fsched-dep-count-heuristic
6559 Enable the dependent-count heuristic in the scheduler. This heuristic
6560 favors the instruction that has more instructions depending on it.
6561 This is enabled by default when scheduling is enabled, i.e.@:
6562 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6563 at @option{-O2} or higher.
6565 @item -freschedule-modulo-scheduled-loops
6566 @opindex freschedule-modulo-scheduled-loops
6567 The modulo scheduling comes before the traditional scheduling, if a loop
6568 was modulo scheduled we may want to prevent the later scheduling passes
6569 from changing its schedule, we use this option to control that.
6571 @item -fselective-scheduling
6572 @opindex fselective-scheduling
6573 Schedule instructions using selective scheduling algorithm. Selective
6574 scheduling runs instead of the first scheduler pass.
6576 @item -fselective-scheduling2
6577 @opindex fselective-scheduling2
6578 Schedule instructions using selective scheduling algorithm. Selective
6579 scheduling runs instead of the second scheduler pass.
6581 @item -fsel-sched-pipelining
6582 @opindex fsel-sched-pipelining
6583 Enable software pipelining of innermost loops during selective scheduling.
6584 This option has no effect until one of @option{-fselective-scheduling} or
6585 @option{-fselective-scheduling2} is turned on.
6587 @item -fsel-sched-pipelining-outer-loops
6588 @opindex fsel-sched-pipelining-outer-loops
6589 When pipelining loops during selective scheduling, also pipeline outer loops.
6590 This option has no effect until @option{-fsel-sched-pipelining} is turned on.
6592 @item -fcaller-saves
6593 @opindex fcaller-saves
6594 Enable values to be allocated in registers that will be clobbered by
6595 function calls, by emitting extra instructions to save and restore the
6596 registers around such calls. Such allocation is done only when it
6597 seems to result in better code than would otherwise be produced.
6599 This option is always enabled by default on certain machines, usually
6600 those which have no call-preserved registers to use instead.
6602 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6604 @item -fconserve-stack
6605 @opindex fconserve-stack
6606 Attempt to minimize stack usage. The compiler will attempt to use less
6607 stack space, even if that makes the program slower. This option
6608 implies setting the @option{large-stack-frame} parameter to 100
6609 and the @option{large-stack-frame-growth} parameter to 400.
6611 @item -ftree-reassoc
6612 @opindex ftree-reassoc
6613 Perform reassociation on trees. This flag is enabled by default
6614 at @option{-O} and higher.
6618 Perform partial redundancy elimination (PRE) on trees. This flag is
6619 enabled by default at @option{-O2} and @option{-O3}.
6621 @item -ftree-forwprop
6622 @opindex ftree-forwprop
6623 Perform forward propagation on trees. This flag is enabled by default
6624 at @option{-O} and higher.
6628 Perform full redundancy elimination (FRE) on trees. The difference
6629 between FRE and PRE is that FRE only considers expressions
6630 that are computed on all paths leading to the redundant computation.
6631 This analysis is faster than PRE, though it exposes fewer redundancies.
6632 This flag is enabled by default at @option{-O} and higher.
6634 @item -ftree-phiprop
6635 @opindex ftree-phiprop
6636 Perform hoisting of loads from conditional pointers on trees. This
6637 pass is enabled by default at @option{-O} and higher.
6639 @item -ftree-copy-prop
6640 @opindex ftree-copy-prop
6641 Perform copy propagation on trees. This pass eliminates unnecessary
6642 copy operations. This flag is enabled by default at @option{-O} and
6645 @item -fipa-pure-const
6646 @opindex fipa-pure-const
6647 Discover which functions are pure or constant.
6648 Enabled by default at @option{-O} and higher.
6650 @item -fipa-reference
6651 @opindex fipa-reference
6652 Discover which static variables do not escape cannot escape the
6654 Enabled by default at @option{-O} and higher.
6656 @item -fipa-struct-reorg
6657 @opindex fipa-struct-reorg
6658 Perform structure reorganization optimization, that change C-like structures
6659 layout in order to better utilize spatial locality. This transformation is
6660 affective for programs containing arrays of structures. Available in two
6661 compilation modes: profile-based (enabled with @option{-fprofile-generate})
6662 or static (which uses built-in heuristics). It works only in whole program
6663 mode, so it requires @option{-fwhole-program} and @option{-combine} to be
6664 enabled. Structures considered @samp{cold} by this transformation are not
6665 affected (see @option{--param struct-reorg-cold-struct-ratio=@var{value}}).
6667 With this flag, the program debug info reflects a new structure layout.
6671 Perform interprocedural pointer analysis and interprocedural modification
6672 and reference analysis. This option can cause excessive memory and
6673 compile-time usage on large compilation units. It is not enabled by
6674 default at any optimization level.
6677 @opindex fipa-profile
6678 Perform interprocedural profile propagation. The functions called only from
6679 cold functions are marked as cold. Also functions executed once (such as
6680 @code{cold}, @code{noreturn}, static constructors or destructors) are identified. Cold
6681 functions and loop less parts of functions executed once are then optimized for
6683 Enabled by default at @option{-O} and higher.
6687 Perform interprocedural constant propagation.
6688 This optimization analyzes the program to determine when values passed
6689 to functions are constants and then optimizes accordingly.
6690 This optimization can substantially increase performance
6691 if the application has constants passed to functions.
6692 This flag is enabled by default at @option{-O2}, @option{-Os} and @option{-O3}.
6694 @item -fipa-cp-clone
6695 @opindex fipa-cp-clone
6696 Perform function cloning to make interprocedural constant propagation stronger.
6697 When enabled, interprocedural constant propagation will perform function cloning
6698 when externally visible function can be called with constant arguments.
6699 Because this optimization can create multiple copies of functions,
6700 it may significantly increase code size
6701 (see @option{--param ipcp-unit-growth=@var{value}}).
6702 This flag is enabled by default at @option{-O3}.
6704 @item -fipa-matrix-reorg
6705 @opindex fipa-matrix-reorg
6706 Perform matrix flattening and transposing.
6707 Matrix flattening tries to replace an @math{m}-dimensional matrix
6708 with its equivalent @math{n}-dimensional matrix, where @math{n < m}.
6709 This reduces the level of indirection needed for accessing the elements
6710 of the matrix. The second optimization is matrix transposing that
6711 attempts to change the order of the matrix's dimensions in order to
6712 improve cache locality.
6713 Both optimizations need the @option{-fwhole-program} flag.
6714 Transposing is enabled only if profiling information is available.
6718 Perform forward store motion on trees. This flag is
6719 enabled by default at @option{-O} and higher.
6723 Perform sparse conditional constant propagation (CCP) on trees. This
6724 pass only operates on local scalar variables and is enabled by default
6725 at @option{-O} and higher.
6727 @item -ftree-switch-conversion
6728 Perform conversion of simple initializations in a switch to
6729 initializations from a scalar array. This flag is enabled by default
6730 at @option{-O2} and higher.
6734 Perform dead code elimination (DCE) on trees. This flag is enabled by
6735 default at @option{-O} and higher.
6737 @item -ftree-builtin-call-dce
6738 @opindex ftree-builtin-call-dce
6739 Perform conditional dead code elimination (DCE) for calls to builtin functions
6740 that may set @code{errno} but are otherwise side-effect free. This flag is
6741 enabled by default at @option{-O2} and higher if @option{-Os} is not also
6744 @item -ftree-dominator-opts
6745 @opindex ftree-dominator-opts
6746 Perform a variety of simple scalar cleanups (constant/copy
6747 propagation, redundancy elimination, range propagation and expression
6748 simplification) based on a dominator tree traversal. This also
6749 performs jump threading (to reduce jumps to jumps). This flag is
6750 enabled by default at @option{-O} and higher.
6754 Perform dead store elimination (DSE) on trees. A dead store is a store into
6755 a memory location which will later be overwritten by another store without
6756 any intervening loads. In this case the earlier store can be deleted. This
6757 flag is enabled by default at @option{-O} and higher.
6761 Perform loop header copying on trees. This is beneficial since it increases
6762 effectiveness of code motion optimizations. It also saves one jump. This flag
6763 is enabled by default at @option{-O} and higher. It is not enabled
6764 for @option{-Os}, since it usually increases code size.
6766 @item -ftree-loop-optimize
6767 @opindex ftree-loop-optimize
6768 Perform loop optimizations on trees. This flag is enabled by default
6769 at @option{-O} and higher.
6771 @item -ftree-loop-linear
6772 @opindex ftree-loop-linear
6773 Perform linear loop transformations on tree. This flag can improve cache
6774 performance and allow further loop optimizations to take place.
6776 @item -floop-interchange
6777 Perform loop interchange transformations on loops. Interchanging two
6778 nested loops switches the inner and outer loops. For example, given a
6783 A(J, I) = A(J, I) * C
6787 loop interchange will transform the loop as if the user had written:
6791 A(J, I) = A(J, I) * C
6795 which can be beneficial when @code{N} is larger than the caches,
6796 because in Fortran, the elements of an array are stored in memory
6797 contiguously by column, and the original loop iterates over rows,
6798 potentially creating at each access a cache miss. This optimization
6799 applies to all the languages supported by GCC and is not limited to
6800 Fortran. To use this code transformation, GCC has to be configured
6801 with @option{--with-ppl} and @option{--with-cloog} to enable the
6802 Graphite loop transformation infrastructure.
6804 @item -floop-strip-mine
6805 Perform loop strip mining transformations on loops. Strip mining
6806 splits a loop into two nested loops. The outer loop has strides
6807 equal to the strip size and the inner loop has strides of the
6808 original loop within a strip. The strip length can be changed
6809 using the @option{loop-block-tile-size} parameter. For example,
6816 loop strip mining will transform the loop as if the user had written:
6819 DO I = II, min (II + 50, N)
6824 This optimization applies to all the languages supported by GCC and is
6825 not limited to Fortran. To use this code transformation, GCC has to
6826 be configured with @option{--with-ppl} and @option{--with-cloog} to
6827 enable the Graphite loop transformation infrastructure.
6830 Perform loop blocking transformations on loops. Blocking strip mines
6831 each loop in the loop nest such that the memory accesses of the
6832 element loops fit inside caches. The strip length can be changed
6833 using the @option{loop-block-tile-size} parameter. For example, given
6838 A(J, I) = B(I) + C(J)
6842 loop blocking will transform the loop as if the user had written:
6846 DO I = II, min (II + 50, N)
6847 DO J = JJ, min (JJ + 50, M)
6848 A(J, I) = B(I) + C(J)
6854 which can be beneficial when @code{M} is larger than the caches,
6855 because the innermost loop will iterate over a smaller amount of data
6856 that can be kept in the caches. This optimization applies to all the
6857 languages supported by GCC and is not limited to Fortran. To use this
6858 code transformation, GCC has to be configured with @option{--with-ppl}
6859 and @option{--with-cloog} to enable the Graphite loop transformation
6862 @item -fgraphite-identity
6863 @opindex fgraphite-identity
6864 Enable the identity transformation for graphite. For every SCoP we generate
6865 the polyhedral representation and transform it back to gimple. Using
6866 @option{-fgraphite-identity} we can check the costs or benefits of the
6867 GIMPLE -> GRAPHITE -> GIMPLE transformation. Some minimal optimizations
6868 are also performed by the code generator CLooG, like index splitting and
6869 dead code elimination in loops.
6871 @item -floop-parallelize-all
6872 Use the Graphite data dependence analysis to identify loops that can
6873 be parallelized. Parallelize all the loops that can be analyzed to
6874 not contain loop carried dependences without checking that it is
6875 profitable to parallelize the loops.
6877 @item -fcheck-data-deps
6878 @opindex fcheck-data-deps
6879 Compare the results of several data dependence analyzers. This option
6880 is used for debugging the data dependence analyzers.
6882 @item -ftree-loop-distribution
6883 Perform loop distribution. This flag can improve cache performance on
6884 big loop bodies and allow further loop optimizations, like
6885 parallelization or vectorization, to take place. For example, the loop
6902 @item -ftree-loop-im
6903 @opindex ftree-loop-im
6904 Perform loop invariant motion on trees. This pass moves only invariants that
6905 would be hard to handle at RTL level (function calls, operations that expand to
6906 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
6907 operands of conditions that are invariant out of the loop, so that we can use
6908 just trivial invariantness analysis in loop unswitching. The pass also includes
6911 @item -ftree-loop-ivcanon
6912 @opindex ftree-loop-ivcanon
6913 Create a canonical counter for number of iterations in the loop for that
6914 determining number of iterations requires complicated analysis. Later
6915 optimizations then may determine the number easily. Useful especially
6916 in connection with unrolling.
6920 Perform induction variable optimizations (strength reduction, induction
6921 variable merging and induction variable elimination) on trees.
6923 @item -ftree-parallelize-loops=n
6924 @opindex ftree-parallelize-loops
6925 Parallelize loops, i.e., split their iteration space to run in n threads.
6926 This is only possible for loops whose iterations are independent
6927 and can be arbitrarily reordered. The optimization is only
6928 profitable on multiprocessor machines, for loops that are CPU-intensive,
6929 rather than constrained e.g.@: by memory bandwidth. This option
6930 implies @option{-pthread}, and thus is only supported on targets
6931 that have support for @option{-pthread}.
6935 Perform function-local points-to analysis on trees. This flag is
6936 enabled by default at @option{-O} and higher.
6940 Perform scalar replacement of aggregates. This pass replaces structure
6941 references with scalars to prevent committing structures to memory too
6942 early. This flag is enabled by default at @option{-O} and higher.
6944 @item -ftree-copyrename
6945 @opindex ftree-copyrename
6946 Perform copy renaming on trees. This pass attempts to rename compiler
6947 temporaries to other variables at copy locations, usually resulting in
6948 variable names which more closely resemble the original variables. This flag
6949 is enabled by default at @option{-O} and higher.
6953 Perform temporary expression replacement during the SSA->normal phase. Single
6954 use/single def temporaries are replaced at their use location with their
6955 defining expression. This results in non-GIMPLE code, but gives the expanders
6956 much more complex trees to work on resulting in better RTL generation. This is
6957 enabled by default at @option{-O} and higher.
6959 @item -ftree-vectorize
6960 @opindex ftree-vectorize
6961 Perform loop vectorization on trees. This flag is enabled by default at
6964 @item -ftree-slp-vectorize
6965 @opindex ftree-slp-vectorize
6966 Perform basic block vectorization on trees. This flag is enabled by default at
6967 @option{-O3} and when @option{-ftree-vectorize} is enabled.
6969 @item -ftree-vect-loop-version
6970 @opindex ftree-vect-loop-version
6971 Perform loop versioning when doing loop vectorization on trees. When a loop
6972 appears to be vectorizable except that data alignment or data dependence cannot
6973 be determined at compile time then vectorized and non-vectorized versions of
6974 the loop are generated along with runtime checks for alignment or dependence
6975 to control which version is executed. This option is enabled by default
6976 except at level @option{-Os} where it is disabled.
6978 @item -fvect-cost-model
6979 @opindex fvect-cost-model
6980 Enable cost model for vectorization.
6984 Perform Value Range Propagation on trees. This is similar to the
6985 constant propagation pass, but instead of values, ranges of values are
6986 propagated. This allows the optimizers to remove unnecessary range
6987 checks like array bound checks and null pointer checks. This is
6988 enabled by default at @option{-O2} and higher. Null pointer check
6989 elimination is only done if @option{-fdelete-null-pointer-checks} is
6994 Perform tail duplication to enlarge superblock size. This transformation
6995 simplifies the control flow of the function allowing other optimizations to do
6998 @item -funroll-loops
6999 @opindex funroll-loops
7000 Unroll loops whose number of iterations can be determined at compile
7001 time or upon entry to the loop. @option{-funroll-loops} implies
7002 @option{-frerun-cse-after-loop}. This option makes code larger,
7003 and may or may not make it run faster.
7005 @item -funroll-all-loops
7006 @opindex funroll-all-loops
7007 Unroll all loops, even if their number of iterations is uncertain when
7008 the loop is entered. This usually makes programs run more slowly.
7009 @option{-funroll-all-loops} implies the same options as
7010 @option{-funroll-loops},
7012 @item -fsplit-ivs-in-unroller
7013 @opindex fsplit-ivs-in-unroller
7014 Enables expressing of values of induction variables in later iterations
7015 of the unrolled loop using the value in the first iteration. This breaks
7016 long dependency chains, thus improving efficiency of the scheduling passes.
7018 Combination of @option{-fweb} and CSE is often sufficient to obtain the
7019 same effect. However in cases the loop body is more complicated than
7020 a single basic block, this is not reliable. It also does not work at all
7021 on some of the architectures due to restrictions in the CSE pass.
7023 This optimization is enabled by default.
7025 @item -fvariable-expansion-in-unroller
7026 @opindex fvariable-expansion-in-unroller
7027 With this option, the compiler will create multiple copies of some
7028 local variables when unrolling a loop which can result in superior code.
7030 @item -fpredictive-commoning
7031 @opindex fpredictive-commoning
7032 Perform predictive commoning optimization, i.e., reusing computations
7033 (especially memory loads and stores) performed in previous
7034 iterations of loops.
7036 This option is enabled at level @option{-O3}.
7038 @item -fprefetch-loop-arrays
7039 @opindex fprefetch-loop-arrays
7040 If supported by the target machine, generate instructions to prefetch
7041 memory to improve the performance of loops that access large arrays.
7043 This option may generate better or worse code; results are highly
7044 dependent on the structure of loops within the source code.
7046 Disabled at level @option{-Os}.
7049 @itemx -fno-peephole2
7050 @opindex fno-peephole
7051 @opindex fno-peephole2
7052 Disable any machine-specific peephole optimizations. The difference
7053 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
7054 are implemented in the compiler; some targets use one, some use the
7055 other, a few use both.
7057 @option{-fpeephole} is enabled by default.
7058 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7060 @item -fno-guess-branch-probability
7061 @opindex fno-guess-branch-probability
7062 Do not guess branch probabilities using heuristics.
7064 GCC will use heuristics to guess branch probabilities if they are
7065 not provided by profiling feedback (@option{-fprofile-arcs}). These
7066 heuristics are based on the control flow graph. If some branch probabilities
7067 are specified by @samp{__builtin_expect}, then the heuristics will be
7068 used to guess branch probabilities for the rest of the control flow graph,
7069 taking the @samp{__builtin_expect} info into account. The interactions
7070 between the heuristics and @samp{__builtin_expect} can be complex, and in
7071 some cases, it may be useful to disable the heuristics so that the effects
7072 of @samp{__builtin_expect} are easier to understand.
7074 The default is @option{-fguess-branch-probability} at levels
7075 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7077 @item -freorder-blocks
7078 @opindex freorder-blocks
7079 Reorder basic blocks in the compiled function in order to reduce number of
7080 taken branches and improve code locality.
7082 Enabled at levels @option{-O2}, @option{-O3}.
7084 @item -freorder-blocks-and-partition
7085 @opindex freorder-blocks-and-partition
7086 In addition to reordering basic blocks in the compiled function, in order
7087 to reduce number of taken branches, partitions hot and cold basic blocks
7088 into separate sections of the assembly and .o files, to improve
7089 paging and cache locality performance.
7091 This optimization is automatically turned off in the presence of
7092 exception handling, for linkonce sections, for functions with a user-defined
7093 section attribute and on any architecture that does not support named
7096 @item -freorder-functions
7097 @opindex freorder-functions
7098 Reorder functions in the object file in order to
7099 improve code locality. This is implemented by using special
7100 subsections @code{.text.hot} for most frequently executed functions and
7101 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
7102 the linker so object file format must support named sections and linker must
7103 place them in a reasonable way.
7105 Also profile feedback must be available in to make this option effective. See
7106 @option{-fprofile-arcs} for details.
7108 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7110 @item -fstrict-aliasing
7111 @opindex fstrict-aliasing
7112 Allow the compiler to assume the strictest aliasing rules applicable to
7113 the language being compiled. For C (and C++), this activates
7114 optimizations based on the type of expressions. In particular, an
7115 object of one type is assumed never to reside at the same address as an
7116 object of a different type, unless the types are almost the same. For
7117 example, an @code{unsigned int} can alias an @code{int}, but not a
7118 @code{void*} or a @code{double}. A character type may alias any other
7121 @anchor{Type-punning}Pay special attention to code like this:
7134 The practice of reading from a different union member than the one most
7135 recently written to (called ``type-punning'') is common. Even with
7136 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
7137 is accessed through the union type. So, the code above will work as
7138 expected. @xref{Structures unions enumerations and bit-fields
7139 implementation}. However, this code might not:
7150 Similarly, access by taking the address, casting the resulting pointer
7151 and dereferencing the result has undefined behavior, even if the cast
7152 uses a union type, e.g.:
7156 return ((union a_union *) &d)->i;
7160 The @option{-fstrict-aliasing} option is enabled at levels
7161 @option{-O2}, @option{-O3}, @option{-Os}.
7163 @item -fstrict-overflow
7164 @opindex fstrict-overflow
7165 Allow the compiler to assume strict signed overflow rules, depending
7166 on the language being compiled. For C (and C++) this means that
7167 overflow when doing arithmetic with signed numbers is undefined, which
7168 means that the compiler may assume that it will not happen. This
7169 permits various optimizations. For example, the compiler will assume
7170 that an expression like @code{i + 10 > i} will always be true for
7171 signed @code{i}. This assumption is only valid if signed overflow is
7172 undefined, as the expression is false if @code{i + 10} overflows when
7173 using twos complement arithmetic. When this option is in effect any
7174 attempt to determine whether an operation on signed numbers will
7175 overflow must be written carefully to not actually involve overflow.
7177 This option also allows the compiler to assume strict pointer
7178 semantics: given a pointer to an object, if adding an offset to that
7179 pointer does not produce a pointer to the same object, the addition is
7180 undefined. This permits the compiler to conclude that @code{p + u >
7181 p} is always true for a pointer @code{p} and unsigned integer
7182 @code{u}. This assumption is only valid because pointer wraparound is
7183 undefined, as the expression is false if @code{p + u} overflows using
7184 twos complement arithmetic.
7186 See also the @option{-fwrapv} option. Using @option{-fwrapv} means
7187 that integer signed overflow is fully defined: it wraps. When
7188 @option{-fwrapv} is used, there is no difference between
7189 @option{-fstrict-overflow} and @option{-fno-strict-overflow} for
7190 integers. With @option{-fwrapv} certain types of overflow are
7191 permitted. For example, if the compiler gets an overflow when doing
7192 arithmetic on constants, the overflowed value can still be used with
7193 @option{-fwrapv}, but not otherwise.
7195 The @option{-fstrict-overflow} option is enabled at levels
7196 @option{-O2}, @option{-O3}, @option{-Os}.
7198 @item -falign-functions
7199 @itemx -falign-functions=@var{n}
7200 @opindex falign-functions
7201 Align the start of functions to the next power-of-two greater than
7202 @var{n}, skipping up to @var{n} bytes. For instance,
7203 @option{-falign-functions=32} aligns functions to the next 32-byte
7204 boundary, but @option{-falign-functions=24} would align to the next
7205 32-byte boundary only if this can be done by skipping 23 bytes or less.
7207 @option{-fno-align-functions} and @option{-falign-functions=1} are
7208 equivalent and mean that functions will not be aligned.
7210 Some assemblers only support this flag when @var{n} is a power of two;
7211 in that case, it is rounded up.
7213 If @var{n} is not specified or is zero, use a machine-dependent default.
7215 Enabled at levels @option{-O2}, @option{-O3}.
7217 @item -falign-labels
7218 @itemx -falign-labels=@var{n}
7219 @opindex falign-labels
7220 Align all branch targets to a power-of-two boundary, skipping up to
7221 @var{n} bytes like @option{-falign-functions}. This option can easily
7222 make code slower, because it must insert dummy operations for when the
7223 branch target is reached in the usual flow of the code.
7225 @option{-fno-align-labels} and @option{-falign-labels=1} are
7226 equivalent and mean that labels will not be aligned.
7228 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
7229 are greater than this value, then their values are used instead.
7231 If @var{n} is not specified or is zero, use a machine-dependent default
7232 which is very likely to be @samp{1}, meaning no alignment.
7234 Enabled at levels @option{-O2}, @option{-O3}.
7237 @itemx -falign-loops=@var{n}
7238 @opindex falign-loops
7239 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
7240 like @option{-falign-functions}. The hope is that the loop will be
7241 executed many times, which will make up for any execution of the dummy
7244 @option{-fno-align-loops} and @option{-falign-loops=1} are
7245 equivalent and mean that loops will not be aligned.
7247 If @var{n} is not specified or is zero, use a machine-dependent default.
7249 Enabled at levels @option{-O2}, @option{-O3}.
7252 @itemx -falign-jumps=@var{n}
7253 @opindex falign-jumps
7254 Align branch targets to a power-of-two boundary, for branch targets
7255 where the targets can only be reached by jumping, skipping up to @var{n}
7256 bytes like @option{-falign-functions}. In this case, no dummy operations
7259 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
7260 equivalent and mean that loops will not be aligned.
7262 If @var{n} is not specified or is zero, use a machine-dependent default.
7264 Enabled at levels @option{-O2}, @option{-O3}.
7266 @item -funit-at-a-time
7267 @opindex funit-at-a-time
7268 This option is left for compatibility reasons. @option{-funit-at-a-time}
7269 has no effect, while @option{-fno-unit-at-a-time} implies
7270 @option{-fno-toplevel-reorder} and @option{-fno-section-anchors}.
7274 @item -fno-toplevel-reorder
7275 @opindex fno-toplevel-reorder
7276 Do not reorder top-level functions, variables, and @code{asm}
7277 statements. Output them in the same order that they appear in the
7278 input file. When this option is used, unreferenced static variables
7279 will not be removed. This option is intended to support existing code
7280 which relies on a particular ordering. For new code, it is better to
7283 Enabled at level @option{-O0}. When disabled explicitly, it also imply
7284 @option{-fno-section-anchors} that is otherwise enabled at @option{-O0} on some
7289 Constructs webs as commonly used for register allocation purposes and assign
7290 each web individual pseudo register. This allows the register allocation pass
7291 to operate on pseudos directly, but also strengthens several other optimization
7292 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
7293 however, make debugging impossible, since variables will no longer stay in a
7296 Enabled by default with @option{-funroll-loops}.
7298 @item -fwhole-program
7299 @opindex fwhole-program
7300 Assume that the current compilation unit represents the whole program being
7301 compiled. All public functions and variables with the exception of @code{main}
7302 and those merged by attribute @code{externally_visible} become static functions
7303 and in effect are optimized more aggressively by interprocedural optimizers.
7304 While this option is equivalent to proper use of the @code{static} keyword for
7305 programs consisting of a single file, in combination with option
7306 @option{-combine}, @option{-flto} or @option{-fwhopr} this flag can be used to
7307 compile many smaller scale programs since the functions and variables become
7308 local for the whole combined compilation unit, not for the single source file
7311 This option implies @option{-fwhole-file} for Fortran programs.
7315 This option runs the standard link-time optimizer. When invoked
7316 with source code, it generates GIMPLE (one of GCC's internal
7317 representations) and writes it to special ELF sections in the object
7318 file. When the object files are linked together, all the function
7319 bodies are read from these ELF sections and instantiated as if they
7320 had been part of the same translation unit.
7322 To use the link-timer optimizer, @option{-flto} needs to be specified at
7323 compile time and during the final link. For example,
7326 gcc -c -O2 -flto foo.c
7327 gcc -c -O2 -flto bar.c
7328 gcc -o myprog -flto -O2 foo.o bar.o
7331 The first two invocations to GCC will save a bytecode representation
7332 of GIMPLE into special ELF sections inside @file{foo.o} and
7333 @file{bar.o}. The final invocation will read the GIMPLE bytecode from
7334 @file{foo.o} and @file{bar.o}, merge the two files into a single
7335 internal image, and compile the result as usual. Since both
7336 @file{foo.o} and @file{bar.o} are merged into a single image, this
7337 causes all the inter-procedural analyses and optimizations in GCC to
7338 work across the two files as if they were a single one. This means,
7339 for example, that the inliner will be able to inline functions in
7340 @file{bar.o} into functions in @file{foo.o} and vice-versa.
7342 Another (simpler) way to enable link-time optimization is,
7345 gcc -o myprog -flto -O2 foo.c bar.c
7348 The above will generate bytecode for @file{foo.c} and @file{bar.c},
7349 merge them together into a single GIMPLE representation and optimize
7350 them as usual to produce @file{myprog}.
7352 The only important thing to keep in mind is that to enable link-time
7353 optimizations the @option{-flto} flag needs to be passed to both the
7354 compile and the link commands.
7356 Note that when a file is compiled with @option{-flto}, the generated
7357 object file will be larger than a regular object file because it will
7358 contain GIMPLE bytecodes and the usual final code. This means that
7359 object files with LTO information can be linked as a normal object
7360 file. So, in the previous example, if the final link is done with
7363 gcc -o myprog foo.o bar.o
7366 The only difference will be that no inter-procedural optimizations
7367 will be applied to produce @file{myprog}. The two object files
7368 @file{foo.o} and @file{bar.o} will be simply sent to the regular
7371 Additionally, the optimization flags used to compile individual files
7372 are not necessarily related to those used at link-time. For instance,
7375 gcc -c -O0 -flto foo.c
7376 gcc -c -O0 -flto bar.c
7377 gcc -o myprog -flto -O3 foo.o bar.o
7380 This will produce individual object files with unoptimized assembler
7381 code, but the resulting binary @file{myprog} will be optimized at
7382 @option{-O3}. Now, if the final binary is generated without
7383 @option{-flto}, then @file{myprog} will not be optimized.
7385 When producing the final binary with @option{-flto}, GCC will only
7386 apply link-time optimizations to those files that contain bytecode.
7387 Therefore, you can mix and match object files and libraries with
7388 GIMPLE bytecodes and final object code. GCC will automatically select
7389 which files to optimize in LTO mode and which files to link without
7392 There are some code generation flags that GCC will preserve when
7393 generating bytecodes, as they need to be used during the final link
7394 stage. Currently, the following options are saved into the GIMPLE
7395 bytecode files: @option{-fPIC}, @option{-fcommon} and all the
7396 @option{-m} target flags.
7398 At link time, these options are read-in and reapplied. Note that the
7399 current implementation makes no attempt at recognizing conflicting
7400 values for these options. If two or more files have a conflicting
7401 value (e.g., one file is compiled with @option{-fPIC} and another
7402 isn't), the compiler will simply use the last value read from the
7403 bytecode files. It is recommended, then, that all the files
7404 participating in the same link be compiled with the same options.
7406 Another feature of LTO is that it is possible to apply interprocedural
7407 optimizations on files written in different languages. This requires
7408 some support in the language front end. Currently, the C, C++ and
7409 Fortran front ends are capable of emitting GIMPLE bytecodes, so
7410 something like this should work
7415 gfortran -c -flto baz.f90
7416 g++ -o myprog -flto -O3 foo.o bar.o baz.o -lgfortran
7419 Notice that the final link is done with @command{g++} to get the C++
7420 runtime libraries and @option{-lgfortran} is added to get the Fortran
7421 runtime libraries. In general, when mixing languages in LTO mode, you
7422 should use the same link command used when mixing languages in a
7423 regular (non-LTO) compilation. This means that if your build process
7424 was mixing languages before, all you need to add is @option{-flto} to
7425 all the compile and link commands.
7427 If LTO encounters objects with C linkage declared with incompatible
7428 types in separate translation units to be linked together (undefined
7429 behavior according to ISO C99 6.2.7), a non-fatal diagnostic may be
7430 issued. The behavior is still undefined at runtime.
7432 If object files containing GIMPLE bytecode are stored in a library
7433 archive, say @file{libfoo.a}, it is possible to extract and use them
7434 in an LTO link if you are using @command{gold} as the linker (which,
7435 in turn requires GCC to be configured with @option{--enable-gold}).
7436 To enable this feature, use the flag @option{-fuse-linker-plugin} at
7440 gcc -o myprog -O2 -flto -fuse-linker-plugin a.o b.o -lfoo
7443 With the linker plugin enabled, @command{gold} will extract the needed
7444 GIMPLE files from @file{libfoo.a} and pass them on to the running GCC
7445 to make them part of the aggregated GIMPLE image to be optimized.
7447 If you are not using @command{gold} and/or do not specify
7448 @option{-fuse-linker-plugin} then the objects inside @file{libfoo.a}
7449 will be extracted and linked as usual, but they will not participate
7450 in the LTO optimization process.
7452 Link time optimizations do not require the presence of the whole
7453 program to operate. If the program does not require any symbols to
7454 be exported, it is possible to combine @option{-flto} and
7455 @option{-fwhopr} with @option{-fwhole-program} to allow the
7456 interprocedural optimizers to use more aggressive assumptions which
7457 may lead to improved optimization opportunities.
7459 Regarding portability: the current implementation of LTO makes no
7460 attempt at generating bytecode that can be ported between different
7461 types of hosts. The bytecode files are versioned and there is a
7462 strict version check, so bytecode files generated in one version of
7463 GCC will not work with an older/newer version of GCC.
7465 Link time optimization does not play well with generating debugging
7466 information. Combining @option{-flto} or @option{-fwhopr} with
7467 @option{-g} is experimental.
7469 This option is disabled by default.
7471 @item -fwhopr[=@var{n}]
7473 This option is identical in functionality to @option{-flto} but it
7474 differs in how the final link stage is executed. Instead of loading
7475 all the function bodies in memory, the callgraph is analyzed and
7476 optimization decisions are made (whole program analysis or WPA). Once
7477 optimization decisions are made, the callgraph is partitioned and the
7478 different sections are compiled separately (local transformations or
7479 LTRANS)@. This process allows optimizations on very large programs
7480 that otherwise would not fit in memory. This option enables
7481 @option{-fwpa} and @option{-fltrans} automatically.
7483 If you specify the optional @var{n} the link stage is executed in
7484 parallel using @var{n} parallel jobs by utilizing an installed
7485 @command{make} program. The environment variable @env{MAKE} may be
7486 used to override the program used.
7488 Disabled by default.
7492 This is an internal option used by GCC when compiling with
7493 @option{-fwhopr}. You should never need to use it.
7495 This option runs the link-time optimizer in the whole-program-analysis
7496 (WPA) mode, which reads in summary information from all inputs and
7497 performs a whole-program analysis based on summary information only.
7498 It generates object files for subsequent runs of the link-time
7499 optimizer where individual object files are optimized using both
7500 summary information from the WPA mode and the actual function bodies.
7501 It then drives the LTRANS phase.
7503 Disabled by default.
7507 This is an internal option used by GCC when compiling with
7508 @option{-fwhopr}. You should never need to use it.
7510 This option runs the link-time optimizer in the local-transformation (LTRANS)
7511 mode, which reads in output from a previous run of the LTO in WPA mode.
7512 In the LTRANS mode, LTO optimizes an object and produces the final assembly.
7514 Disabled by default.
7516 @item -fltrans-output-list=@var{file}
7517 @opindex fltrans-output-list
7518 This is an internal option used by GCC when compiling with
7519 @option{-fwhopr}. You should never need to use it.
7521 This option specifies a file to which the names of LTRANS output files are
7522 written. This option is only meaningful in conjunction with @option{-fwpa}.
7524 Disabled by default.
7526 @item -flto-compression-level=@var{n}
7527 This option specifies the level of compression used for intermediate
7528 language written to LTO object files, and is only meaningful in
7529 conjunction with LTO mode (@option{-fwhopr}, @option{-flto}). Valid
7530 values are 0 (no compression) to 9 (maximum compression). Values
7531 outside this range are clamped to either 0 or 9. If the option is not
7532 given, a default balanced compression setting is used.
7535 Prints a report with internal details on the workings of the link-time
7536 optimizer. The contents of this report vary from version to version,
7537 it is meant to be useful to GCC developers when processing object
7538 files in LTO mode (via @option{-fwhopr} or @option{-flto}).
7540 Disabled by default.
7542 @item -fuse-linker-plugin
7543 Enables the extraction of objects with GIMPLE bytecode information
7544 from library archives. This option relies on features available only
7545 in @command{gold}, so to use this you must configure GCC with
7546 @option{--enable-gold}. See @option{-flto} for a description on the
7547 effect of this flag and how to use it.
7549 Disabled by default.
7551 @item -fcprop-registers
7552 @opindex fcprop-registers
7553 After register allocation and post-register allocation instruction splitting,
7554 we perform a copy-propagation pass to try to reduce scheduling dependencies
7555 and occasionally eliminate the copy.
7557 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7559 @item -fprofile-correction
7560 @opindex fprofile-correction
7561 Profiles collected using an instrumented binary for multi-threaded programs may
7562 be inconsistent due to missed counter updates. When this option is specified,
7563 GCC will use heuristics to correct or smooth out such inconsistencies. By
7564 default, GCC will emit an error message when an inconsistent profile is detected.
7566 @item -fprofile-dir=@var{path}
7567 @opindex fprofile-dir
7569 Set the directory to search the profile data files in to @var{path}.
7570 This option affects only the profile data generated by
7571 @option{-fprofile-generate}, @option{-ftest-coverage}, @option{-fprofile-arcs}
7572 and used by @option{-fprofile-use} and @option{-fbranch-probabilities}
7573 and its related options.
7574 By default, GCC will use the current directory as @var{path}
7575 thus the profile data file will appear in the same directory as the object file.
7577 @item -fprofile-generate
7578 @itemx -fprofile-generate=@var{path}
7579 @opindex fprofile-generate
7581 Enable options usually used for instrumenting application to produce
7582 profile useful for later recompilation with profile feedback based
7583 optimization. You must use @option{-fprofile-generate} both when
7584 compiling and when linking your program.
7586 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
7588 If @var{path} is specified, GCC will look at the @var{path} to find
7589 the profile feedback data files. See @option{-fprofile-dir}.
7592 @itemx -fprofile-use=@var{path}
7593 @opindex fprofile-use
7594 Enable profile feedback directed optimizations, and optimizations
7595 generally profitable only with profile feedback available.
7597 The following options are enabled: @code{-fbranch-probabilities}, @code{-fvpt},
7598 @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}
7600 By default, GCC emits an error message if the feedback profiles do not
7601 match the source code. This error can be turned into a warning by using
7602 @option{-Wcoverage-mismatch}. Note this may result in poorly optimized
7605 If @var{path} is specified, GCC will look at the @var{path} to find
7606 the profile feedback data files. See @option{-fprofile-dir}.
7609 The following options control compiler behavior regarding floating
7610 point arithmetic. These options trade off between speed and
7611 correctness. All must be specifically enabled.
7615 @opindex ffloat-store
7616 Do not store floating point variables in registers, and inhibit other
7617 options that might change whether a floating point value is taken from a
7620 @cindex floating point precision
7621 This option prevents undesirable excess precision on machines such as
7622 the 68000 where the floating registers (of the 68881) keep more
7623 precision than a @code{double} is supposed to have. Similarly for the
7624 x86 architecture. For most programs, the excess precision does only
7625 good, but a few programs rely on the precise definition of IEEE floating
7626 point. Use @option{-ffloat-store} for such programs, after modifying
7627 them to store all pertinent intermediate computations into variables.
7629 @item -fexcess-precision=@var{style}
7630 @opindex fexcess-precision
7631 This option allows further control over excess precision on machines
7632 where floating-point registers have more precision than the IEEE
7633 @code{float} and @code{double} types and the processor does not
7634 support operations rounding to those types. By default,
7635 @option{-fexcess-precision=fast} is in effect; this means that
7636 operations are carried out in the precision of the registers and that
7637 it is unpredictable when rounding to the types specified in the source
7638 code takes place. When compiling C, if
7639 @option{-fexcess-precision=standard} is specified then excess
7640 precision will follow the rules specified in ISO C99; in particular,
7641 both casts and assignments cause values to be rounded to their
7642 semantic types (whereas @option{-ffloat-store} only affects
7643 assignments). This option is enabled by default for C if a strict
7644 conformance option such as @option{-std=c99} is used.
7647 @option{-fexcess-precision=standard} is not implemented for languages
7648 other than C, and has no effect if
7649 @option{-funsafe-math-optimizations} or @option{-ffast-math} is
7650 specified. On the x86, it also has no effect if @option{-mfpmath=sse}
7651 or @option{-mfpmath=sse+387} is specified; in the former case, IEEE
7652 semantics apply without excess precision, and in the latter, rounding
7657 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations},
7658 @option{-ffinite-math-only}, @option{-fno-rounding-math},
7659 @option{-fno-signaling-nans} and @option{-fcx-limited-range}.
7661 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
7663 This option is not turned on by any @option{-O} option since
7664 it can result in incorrect output for programs which depend on
7665 an exact implementation of IEEE or ISO rules/specifications for
7666 math functions. It may, however, yield faster code for programs
7667 that do not require the guarantees of these specifications.
7669 @item -fno-math-errno
7670 @opindex fno-math-errno
7671 Do not set ERRNO after calling math functions that are executed
7672 with a single instruction, e.g., sqrt. A program that relies on
7673 IEEE exceptions for math error handling may want to use this flag
7674 for speed while maintaining IEEE arithmetic compatibility.
7676 This option is not turned on by any @option{-O} option since
7677 it can result in incorrect output for programs which depend on
7678 an exact implementation of IEEE or ISO rules/specifications for
7679 math functions. It may, however, yield faster code for programs
7680 that do not require the guarantees of these specifications.
7682 The default is @option{-fmath-errno}.
7684 On Darwin systems, the math library never sets @code{errno}. There is
7685 therefore no reason for the compiler to consider the possibility that
7686 it might, and @option{-fno-math-errno} is the default.
7688 @item -funsafe-math-optimizations
7689 @opindex funsafe-math-optimizations
7691 Allow optimizations for floating-point arithmetic that (a) assume
7692 that arguments and results are valid and (b) may violate IEEE or
7693 ANSI standards. When used at link-time, it may include libraries
7694 or startup files that change the default FPU control word or other
7695 similar optimizations.
7697 This option is not turned on by any @option{-O} option since
7698 it can result in incorrect output for programs which depend on
7699 an exact implementation of IEEE or ISO rules/specifications for
7700 math functions. It may, however, yield faster code for programs
7701 that do not require the guarantees of these specifications.
7702 Enables @option{-fno-signed-zeros}, @option{-fno-trapping-math},
7703 @option{-fassociative-math} and @option{-freciprocal-math}.
7705 The default is @option{-fno-unsafe-math-optimizations}.
7707 @item -fassociative-math
7708 @opindex fassociative-math
7710 Allow re-association of operands in series of floating-point operations.
7711 This violates the ISO C and C++ language standard by possibly changing
7712 computation result. NOTE: re-ordering may change the sign of zero as
7713 well as ignore NaNs and inhibit or create underflow or overflow (and
7714 thus cannot be used on a code which relies on rounding behavior like
7715 @code{(x + 2**52) - 2**52)}. May also reorder floating-point comparisons
7716 and thus may not be used when ordered comparisons are required.
7717 This option requires that both @option{-fno-signed-zeros} and
7718 @option{-fno-trapping-math} be in effect. Moreover, it doesn't make
7719 much sense with @option{-frounding-math}. For Fortran the option
7720 is automatically enabled when both @option{-fno-signed-zeros} and
7721 @option{-fno-trapping-math} are in effect.
7723 The default is @option{-fno-associative-math}.
7725 @item -freciprocal-math
7726 @opindex freciprocal-math
7728 Allow the reciprocal of a value to be used instead of dividing by
7729 the value if this enables optimizations. For example @code{x / y}
7730 can be replaced with @code{x * (1/y)} which is useful if @code{(1/y)}
7731 is subject to common subexpression elimination. Note that this loses
7732 precision and increases the number of flops operating on the value.
7734 The default is @option{-fno-reciprocal-math}.
7736 @item -ffinite-math-only
7737 @opindex ffinite-math-only
7738 Allow optimizations for floating-point arithmetic that assume
7739 that arguments and results are not NaNs or +-Infs.
7741 This option is not turned on by any @option{-O} option since
7742 it can result in incorrect output for programs which depend on
7743 an exact implementation of IEEE or ISO rules/specifications for
7744 math functions. It may, however, yield faster code for programs
7745 that do not require the guarantees of these specifications.
7747 The default is @option{-fno-finite-math-only}.
7749 @item -fno-signed-zeros
7750 @opindex fno-signed-zeros
7751 Allow optimizations for floating point arithmetic that ignore the
7752 signedness of zero. IEEE arithmetic specifies the behavior of
7753 distinct +0.0 and @minus{}0.0 values, which then prohibits simplification
7754 of expressions such as x+0.0 or 0.0*x (even with @option{-ffinite-math-only}).
7755 This option implies that the sign of a zero result isn't significant.
7757 The default is @option{-fsigned-zeros}.
7759 @item -fno-trapping-math
7760 @opindex fno-trapping-math
7761 Compile code assuming that floating-point operations cannot generate
7762 user-visible traps. These traps include division by zero, overflow,
7763 underflow, inexact result and invalid operation. This option requires
7764 that @option{-fno-signaling-nans} be in effect. Setting this option may
7765 allow faster code if one relies on ``non-stop'' IEEE arithmetic, for example.
7767 This option should never be turned on by any @option{-O} option since
7768 it can result in incorrect output for programs which depend on
7769 an exact implementation of IEEE or ISO rules/specifications for
7772 The default is @option{-ftrapping-math}.
7774 @item -frounding-math
7775 @opindex frounding-math
7776 Disable transformations and optimizations that assume default floating
7777 point rounding behavior. This is round-to-zero for all floating point
7778 to integer conversions, and round-to-nearest for all other arithmetic
7779 truncations. This option should be specified for programs that change
7780 the FP rounding mode dynamically, or that may be executed with a
7781 non-default rounding mode. This option disables constant folding of
7782 floating point expressions at compile-time (which may be affected by
7783 rounding mode) and arithmetic transformations that are unsafe in the
7784 presence of sign-dependent rounding modes.
7786 The default is @option{-fno-rounding-math}.
7788 This option is experimental and does not currently guarantee to
7789 disable all GCC optimizations that are affected by rounding mode.
7790 Future versions of GCC may provide finer control of this setting
7791 using C99's @code{FENV_ACCESS} pragma. This command line option
7792 will be used to specify the default state for @code{FENV_ACCESS}.
7794 @item -fsignaling-nans
7795 @opindex fsignaling-nans
7796 Compile code assuming that IEEE signaling NaNs may generate user-visible
7797 traps during floating-point operations. Setting this option disables
7798 optimizations that may change the number of exceptions visible with
7799 signaling NaNs. This option implies @option{-ftrapping-math}.
7801 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
7804 The default is @option{-fno-signaling-nans}.
7806 This option is experimental and does not currently guarantee to
7807 disable all GCC optimizations that affect signaling NaN behavior.
7809 @item -fsingle-precision-constant
7810 @opindex fsingle-precision-constant
7811 Treat floating point constant as single precision constant instead of
7812 implicitly converting it to double precision constant.
7814 @item -fcx-limited-range
7815 @opindex fcx-limited-range
7816 When enabled, this option states that a range reduction step is not
7817 needed when performing complex division. Also, there is no checking
7818 whether the result of a complex multiplication or division is @code{NaN
7819 + I*NaN}, with an attempt to rescue the situation in that case. The
7820 default is @option{-fno-cx-limited-range}, but is enabled by
7821 @option{-ffast-math}.
7823 This option controls the default setting of the ISO C99
7824 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
7827 @item -fcx-fortran-rules
7828 @opindex fcx-fortran-rules
7829 Complex multiplication and division follow Fortran rules. Range
7830 reduction is done as part of complex division, but there is no checking
7831 whether the result of a complex multiplication or division is @code{NaN
7832 + I*NaN}, with an attempt to rescue the situation in that case.
7834 The default is @option{-fno-cx-fortran-rules}.
7838 The following options control optimizations that may improve
7839 performance, but are not enabled by any @option{-O} options. This
7840 section includes experimental options that may produce broken code.
7843 @item -fbranch-probabilities
7844 @opindex fbranch-probabilities
7845 After running a program compiled with @option{-fprofile-arcs}
7846 (@pxref{Debugging Options,, Options for Debugging Your Program or
7847 @command{gcc}}), you can compile it a second time using
7848 @option{-fbranch-probabilities}, to improve optimizations based on
7849 the number of times each branch was taken. When the program
7850 compiled with @option{-fprofile-arcs} exits it saves arc execution
7851 counts to a file called @file{@var{sourcename}.gcda} for each source
7852 file. The information in this data file is very dependent on the
7853 structure of the generated code, so you must use the same source code
7854 and the same optimization options for both compilations.
7856 With @option{-fbranch-probabilities}, GCC puts a
7857 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
7858 These can be used to improve optimization. Currently, they are only
7859 used in one place: in @file{reorg.c}, instead of guessing which path a
7860 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
7861 exactly determine which path is taken more often.
7863 @item -fprofile-values
7864 @opindex fprofile-values
7865 If combined with @option{-fprofile-arcs}, it adds code so that some
7866 data about values of expressions in the program is gathered.
7868 With @option{-fbranch-probabilities}, it reads back the data gathered
7869 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
7870 notes to instructions for their later usage in optimizations.
7872 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
7876 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
7877 a code to gather information about values of expressions.
7879 With @option{-fbranch-probabilities}, it reads back the data gathered
7880 and actually performs the optimizations based on them.
7881 Currently the optimizations include specialization of division operation
7882 using the knowledge about the value of the denominator.
7884 @item -frename-registers
7885 @opindex frename-registers
7886 Attempt to avoid false dependencies in scheduled code by making use
7887 of registers left over after register allocation. This optimization
7888 will most benefit processors with lots of registers. Depending on the
7889 debug information format adopted by the target, however, it can
7890 make debugging impossible, since variables will no longer stay in
7891 a ``home register''.
7893 Enabled by default with @option{-funroll-loops} and @option{-fpeel-loops}.
7897 Perform tail duplication to enlarge superblock size. This transformation
7898 simplifies the control flow of the function allowing other optimizations to do
7901 Enabled with @option{-fprofile-use}.
7903 @item -funroll-loops
7904 @opindex funroll-loops
7905 Unroll loops whose number of iterations can be determined at compile time or
7906 upon entry to the loop. @option{-funroll-loops} implies
7907 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
7908 It also turns on complete loop peeling (i.e.@: complete removal of loops with
7909 small constant number of iterations). This option makes code larger, and may
7910 or may not make it run faster.
7912 Enabled with @option{-fprofile-use}.
7914 @item -funroll-all-loops
7915 @opindex funroll-all-loops
7916 Unroll all loops, even if their number of iterations is uncertain when
7917 the loop is entered. This usually makes programs run more slowly.
7918 @option{-funroll-all-loops} implies the same options as
7919 @option{-funroll-loops}.
7922 @opindex fpeel-loops
7923 Peels the loops for that there is enough information that they do not
7924 roll much (from profile feedback). It also turns on complete loop peeling
7925 (i.e.@: complete removal of loops with small constant number of iterations).
7927 Enabled with @option{-fprofile-use}.
7929 @item -fmove-loop-invariants
7930 @opindex fmove-loop-invariants
7931 Enables the loop invariant motion pass in the RTL loop optimizer. Enabled
7932 at level @option{-O1}
7934 @item -funswitch-loops
7935 @opindex funswitch-loops
7936 Move branches with loop invariant conditions out of the loop, with duplicates
7937 of the loop on both branches (modified according to result of the condition).
7939 @item -ffunction-sections
7940 @itemx -fdata-sections
7941 @opindex ffunction-sections
7942 @opindex fdata-sections
7943 Place each function or data item into its own section in the output
7944 file if the target supports arbitrary sections. The name of the
7945 function or the name of the data item determines the section's name
7948 Use these options on systems where the linker can perform optimizations
7949 to improve locality of reference in the instruction space. Most systems
7950 using the ELF object format and SPARC processors running Solaris 2 have
7951 linkers with such optimizations. AIX may have these optimizations in
7954 Only use these options when there are significant benefits from doing
7955 so. When you specify these options, the assembler and linker will
7956 create larger object and executable files and will also be slower.
7957 You will not be able to use @code{gprof} on all systems if you
7958 specify this option and you may have problems with debugging if
7959 you specify both this option and @option{-g}.
7961 @item -fbranch-target-load-optimize
7962 @opindex fbranch-target-load-optimize
7963 Perform branch target register load optimization before prologue / epilogue
7965 The use of target registers can typically be exposed only during reload,
7966 thus hoisting loads out of loops and doing inter-block scheduling needs
7967 a separate optimization pass.
7969 @item -fbranch-target-load-optimize2
7970 @opindex fbranch-target-load-optimize2
7971 Perform branch target register load optimization after prologue / epilogue
7974 @item -fbtr-bb-exclusive
7975 @opindex fbtr-bb-exclusive
7976 When performing branch target register load optimization, don't reuse
7977 branch target registers in within any basic block.
7979 @item -fstack-protector
7980 @opindex fstack-protector
7981 Emit extra code to check for buffer overflows, such as stack smashing
7982 attacks. This is done by adding a guard variable to functions with
7983 vulnerable objects. This includes functions that call alloca, and
7984 functions with buffers larger than 8 bytes. The guards are initialized
7985 when a function is entered and then checked when the function exits.
7986 If a guard check fails, an error message is printed and the program exits.
7988 @item -fstack-protector-all
7989 @opindex fstack-protector-all
7990 Like @option{-fstack-protector} except that all functions are protected.
7992 @item -fsection-anchors
7993 @opindex fsection-anchors
7994 Try to reduce the number of symbolic address calculations by using
7995 shared ``anchor'' symbols to address nearby objects. This transformation
7996 can help to reduce the number of GOT entries and GOT accesses on some
7999 For example, the implementation of the following function @code{foo}:
8003 int foo (void) @{ return a + b + c; @}
8006 would usually calculate the addresses of all three variables, but if you
8007 compile it with @option{-fsection-anchors}, it will access the variables
8008 from a common anchor point instead. The effect is similar to the
8009 following pseudocode (which isn't valid C):
8014 register int *xr = &x;
8015 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
8019 Not all targets support this option.
8021 @item --param @var{name}=@var{value}
8023 In some places, GCC uses various constants to control the amount of
8024 optimization that is done. For example, GCC will not inline functions
8025 that contain more that a certain number of instructions. You can
8026 control some of these constants on the command-line using the
8027 @option{--param} option.
8029 The names of specific parameters, and the meaning of the values, are
8030 tied to the internals of the compiler, and are subject to change
8031 without notice in future releases.
8033 In each case, the @var{value} is an integer. The allowable choices for
8034 @var{name} are given in the following table:
8037 @item struct-reorg-cold-struct-ratio
8038 The threshold ratio (as a percentage) between a structure frequency
8039 and the frequency of the hottest structure in the program. This parameter
8040 is used by struct-reorg optimization enabled by @option{-fipa-struct-reorg}.
8041 We say that if the ratio of a structure frequency, calculated by profiling,
8042 to the hottest structure frequency in the program is less than this
8043 parameter, then structure reorganization is not applied to this structure.
8046 @item predictable-branch-outcome
8047 When branch is predicted to be taken with probability lower than this threshold
8048 (in percent), then it is considered well predictable. The default is 10.
8050 @item max-crossjump-edges
8051 The maximum number of incoming edges to consider for crossjumping.
8052 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
8053 the number of edges incoming to each block. Increasing values mean
8054 more aggressive optimization, making the compile time increase with
8055 probably small improvement in executable size.
8057 @item min-crossjump-insns
8058 The minimum number of instructions which must be matched at the end
8059 of two blocks before crossjumping will be performed on them. This
8060 value is ignored in the case where all instructions in the block being
8061 crossjumped from are matched. The default value is 5.
8063 @item max-grow-copy-bb-insns
8064 The maximum code size expansion factor when copying basic blocks
8065 instead of jumping. The expansion is relative to a jump instruction.
8066 The default value is 8.
8068 @item max-goto-duplication-insns
8069 The maximum number of instructions to duplicate to a block that jumps
8070 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
8071 passes, GCC factors computed gotos early in the compilation process,
8072 and unfactors them as late as possible. Only computed jumps at the
8073 end of a basic blocks with no more than max-goto-duplication-insns are
8074 unfactored. The default value is 8.
8076 @item max-delay-slot-insn-search
8077 The maximum number of instructions to consider when looking for an
8078 instruction to fill a delay slot. If more than this arbitrary number of
8079 instructions is searched, the time savings from filling the delay slot
8080 will be minimal so stop searching. Increasing values mean more
8081 aggressive optimization, making the compile time increase with probably
8082 small improvement in executable run time.
8084 @item max-delay-slot-live-search
8085 When trying to fill delay slots, the maximum number of instructions to
8086 consider when searching for a block with valid live register
8087 information. Increasing this arbitrarily chosen value means more
8088 aggressive optimization, increasing the compile time. This parameter
8089 should be removed when the delay slot code is rewritten to maintain the
8092 @item max-gcse-memory
8093 The approximate maximum amount of memory that will be allocated in
8094 order to perform the global common subexpression elimination
8095 optimization. If more memory than specified is required, the
8096 optimization will not be done.
8098 @item max-pending-list-length
8099 The maximum number of pending dependencies scheduling will allow
8100 before flushing the current state and starting over. Large functions
8101 with few branches or calls can create excessively large lists which
8102 needlessly consume memory and resources.
8104 @item max-inline-insns-single
8105 Several parameters control the tree inliner used in gcc.
8106 This number sets the maximum number of instructions (counted in GCC's
8107 internal representation) in a single function that the tree inliner
8108 will consider for inlining. This only affects functions declared
8109 inline and methods implemented in a class declaration (C++).
8110 The default value is 300.
8112 @item max-inline-insns-auto
8113 When you use @option{-finline-functions} (included in @option{-O3}),
8114 a lot of functions that would otherwise not be considered for inlining
8115 by the compiler will be investigated. To those functions, a different
8116 (more restrictive) limit compared to functions declared inline can
8118 The default value is 50.
8120 @item large-function-insns
8121 The limit specifying really large functions. For functions larger than this
8122 limit after inlining, inlining is constrained by
8123 @option{--param large-function-growth}. This parameter is useful primarily
8124 to avoid extreme compilation time caused by non-linear algorithms used by the
8126 The default value is 2700.
8128 @item large-function-growth
8129 Specifies maximal growth of large function caused by inlining in percents.
8130 The default value is 100 which limits large function growth to 2.0 times
8133 @item large-unit-insns
8134 The limit specifying large translation unit. Growth caused by inlining of
8135 units larger than this limit is limited by @option{--param inline-unit-growth}.
8136 For small units this might be too tight (consider unit consisting of function A
8137 that is inline and B that just calls A three time. If B is small relative to
8138 A, the growth of unit is 300\% and yet such inlining is very sane. For very
8139 large units consisting of small inlineable functions however the overall unit
8140 growth limit is needed to avoid exponential explosion of code size. Thus for
8141 smaller units, the size is increased to @option{--param large-unit-insns}
8142 before applying @option{--param inline-unit-growth}. The default is 10000
8144 @item inline-unit-growth
8145 Specifies maximal overall growth of the compilation unit caused by inlining.
8146 The default value is 30 which limits unit growth to 1.3 times the original
8149 @item ipcp-unit-growth
8150 Specifies maximal overall growth of the compilation unit caused by
8151 interprocedural constant propagation. The default value is 10 which limits
8152 unit growth to 1.1 times the original size.
8154 @item large-stack-frame
8155 The limit specifying large stack frames. While inlining the algorithm is trying
8156 to not grow past this limit too much. Default value is 256 bytes.
8158 @item large-stack-frame-growth
8159 Specifies maximal growth of large stack frames caused by inlining in percents.
8160 The default value is 1000 which limits large stack frame growth to 11 times
8163 @item max-inline-insns-recursive
8164 @itemx max-inline-insns-recursive-auto
8165 Specifies maximum number of instructions out-of-line copy of self recursive inline
8166 function can grow into by performing recursive inlining.
8168 For functions declared inline @option{--param max-inline-insns-recursive} is
8169 taken into account. For function not declared inline, recursive inlining
8170 happens only when @option{-finline-functions} (included in @option{-O3}) is
8171 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
8172 default value is 450.
8174 @item max-inline-recursive-depth
8175 @itemx max-inline-recursive-depth-auto
8176 Specifies maximum recursion depth used by the recursive inlining.
8178 For functions declared inline @option{--param max-inline-recursive-depth} is
8179 taken into account. For function not declared inline, recursive inlining
8180 happens only when @option{-finline-functions} (included in @option{-O3}) is
8181 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
8184 @item min-inline-recursive-probability
8185 Recursive inlining is profitable only for function having deep recursion
8186 in average and can hurt for function having little recursion depth by
8187 increasing the prologue size or complexity of function body to other
8190 When profile feedback is available (see @option{-fprofile-generate}) the actual
8191 recursion depth can be guessed from probability that function will recurse via
8192 given call expression. This parameter limits inlining only to call expression
8193 whose probability exceeds given threshold (in percents). The default value is
8196 @item early-inlining-insns
8197 Specify growth that early inliner can make. In effect it increases amount of
8198 inlining for code having large abstraction penalty. The default value is 8.
8200 @item max-early-inliner-iterations
8201 @itemx max-early-inliner-iterations
8202 Limit of iterations of early inliner. This basically bounds number of nested
8203 indirect calls early inliner can resolve. Deeper chains are still handled by
8206 @item min-vect-loop-bound
8207 The minimum number of iterations under which a loop will not get vectorized
8208 when @option{-ftree-vectorize} is used. The number of iterations after
8209 vectorization needs to be greater than the value specified by this option
8210 to allow vectorization. The default value is 0.
8212 @item max-unrolled-insns
8213 The maximum number of instructions that a loop should have if that loop
8214 is unrolled, and if the loop is unrolled, it determines how many times
8215 the loop code is unrolled.
8217 @item max-average-unrolled-insns
8218 The maximum number of instructions biased by probabilities of their execution
8219 that a loop should have if that loop is unrolled, and if the loop is unrolled,
8220 it determines how many times the loop code is unrolled.
8222 @item max-unroll-times
8223 The maximum number of unrollings of a single loop.
8225 @item max-peeled-insns
8226 The maximum number of instructions that a loop should have if that loop
8227 is peeled, and if the loop is peeled, it determines how many times
8228 the loop code is peeled.
8230 @item max-peel-times
8231 The maximum number of peelings of a single loop.
8233 @item max-completely-peeled-insns
8234 The maximum number of insns of a completely peeled loop.
8236 @item max-completely-peel-times
8237 The maximum number of iterations of a loop to be suitable for complete peeling.
8239 @item max-completely-peel-loop-nest-depth
8240 The maximum depth of a loop nest suitable for complete peeling.
8242 @item max-unswitch-insns
8243 The maximum number of insns of an unswitched loop.
8245 @item max-unswitch-level
8246 The maximum number of branches unswitched in a single loop.
8249 The minimum cost of an expensive expression in the loop invariant motion.
8251 @item iv-consider-all-candidates-bound
8252 Bound on number of candidates for induction variables below that
8253 all candidates are considered for each use in induction variable
8254 optimizations. Only the most relevant candidates are considered
8255 if there are more candidates, to avoid quadratic time complexity.
8257 @item iv-max-considered-uses
8258 The induction variable optimizations give up on loops that contain more
8259 induction variable uses.
8261 @item iv-always-prune-cand-set-bound
8262 If number of candidates in the set is smaller than this value,
8263 we always try to remove unnecessary ivs from the set during its
8264 optimization when a new iv is added to the set.
8266 @item scev-max-expr-size
8267 Bound on size of expressions used in the scalar evolutions analyzer.
8268 Large expressions slow the analyzer.
8270 @item omega-max-vars
8271 The maximum number of variables in an Omega constraint system.
8272 The default value is 128.
8274 @item omega-max-geqs
8275 The maximum number of inequalities in an Omega constraint system.
8276 The default value is 256.
8279 The maximum number of equalities in an Omega constraint system.
8280 The default value is 128.
8282 @item omega-max-wild-cards
8283 The maximum number of wildcard variables that the Omega solver will
8284 be able to insert. The default value is 18.
8286 @item omega-hash-table-size
8287 The size of the hash table in the Omega solver. The default value is
8290 @item omega-max-keys
8291 The maximal number of keys used by the Omega solver. The default
8294 @item omega-eliminate-redundant-constraints
8295 When set to 1, use expensive methods to eliminate all redundant
8296 constraints. The default value is 0.
8298 @item vect-max-version-for-alignment-checks
8299 The maximum number of runtime checks that can be performed when
8300 doing loop versioning for alignment in the vectorizer. See option
8301 ftree-vect-loop-version for more information.
8303 @item vect-max-version-for-alias-checks
8304 The maximum number of runtime checks that can be performed when
8305 doing loop versioning for alias in the vectorizer. See option
8306 ftree-vect-loop-version for more information.
8308 @item max-iterations-to-track
8310 The maximum number of iterations of a loop the brute force algorithm
8311 for analysis of # of iterations of the loop tries to evaluate.
8313 @item hot-bb-count-fraction
8314 Select fraction of the maximal count of repetitions of basic block in program
8315 given basic block needs to have to be considered hot.
8317 @item hot-bb-frequency-fraction
8318 Select fraction of the maximal frequency of executions of basic block in
8319 function given basic block needs to have to be considered hot
8321 @item max-predicted-iterations
8322 The maximum number of loop iterations we predict statically. This is useful
8323 in cases where function contain single loop with known bound and other loop
8324 with unknown. We predict the known number of iterations correctly, while
8325 the unknown number of iterations average to roughly 10. This means that the
8326 loop without bounds would appear artificially cold relative to the other one.
8328 @item align-threshold
8330 Select fraction of the maximal frequency of executions of basic block in
8331 function given basic block will get aligned.
8333 @item align-loop-iterations
8335 A loop expected to iterate at lest the selected number of iterations will get
8338 @item tracer-dynamic-coverage
8339 @itemx tracer-dynamic-coverage-feedback
8341 This value is used to limit superblock formation once the given percentage of
8342 executed instructions is covered. This limits unnecessary code size
8345 The @option{tracer-dynamic-coverage-feedback} is used only when profile
8346 feedback is available. The real profiles (as opposed to statically estimated
8347 ones) are much less balanced allowing the threshold to be larger value.
8349 @item tracer-max-code-growth
8350 Stop tail duplication once code growth has reached given percentage. This is
8351 rather hokey argument, as most of the duplicates will be eliminated later in
8352 cross jumping, so it may be set to much higher values than is the desired code
8355 @item tracer-min-branch-ratio
8357 Stop reverse growth when the reverse probability of best edge is less than this
8358 threshold (in percent).
8360 @item tracer-min-branch-ratio
8361 @itemx tracer-min-branch-ratio-feedback
8363 Stop forward growth if the best edge do have probability lower than this
8366 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
8367 compilation for profile feedback and one for compilation without. The value
8368 for compilation with profile feedback needs to be more conservative (higher) in
8369 order to make tracer effective.
8371 @item max-cse-path-length
8373 Maximum number of basic blocks on path that cse considers. The default is 10.
8376 The maximum instructions CSE process before flushing. The default is 1000.
8378 @item ggc-min-expand
8380 GCC uses a garbage collector to manage its own memory allocation. This
8381 parameter specifies the minimum percentage by which the garbage
8382 collector's heap should be allowed to expand between collections.
8383 Tuning this may improve compilation speed; it has no effect on code
8386 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
8387 RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is
8388 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
8389 GCC is not able to calculate RAM on a particular platform, the lower
8390 bound of 30% is used. Setting this parameter and
8391 @option{ggc-min-heapsize} to zero causes a full collection to occur at
8392 every opportunity. This is extremely slow, but can be useful for
8395 @item ggc-min-heapsize
8397 Minimum size of the garbage collector's heap before it begins bothering
8398 to collect garbage. The first collection occurs after the heap expands
8399 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
8400 tuning this may improve compilation speed, and has no effect on code
8403 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which
8404 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
8405 with a lower bound of 4096 (four megabytes) and an upper bound of
8406 131072 (128 megabytes). If GCC is not able to calculate RAM on a
8407 particular platform, the lower bound is used. Setting this parameter
8408 very large effectively disables garbage collection. Setting this
8409 parameter and @option{ggc-min-expand} to zero causes a full collection
8410 to occur at every opportunity.
8412 @item max-reload-search-insns
8413 The maximum number of instruction reload should look backward for equivalent
8414 register. Increasing values mean more aggressive optimization, making the
8415 compile time increase with probably slightly better performance. The default
8418 @item max-cselib-memory-locations
8419 The maximum number of memory locations cselib should take into account.
8420 Increasing values mean more aggressive optimization, making the compile time
8421 increase with probably slightly better performance. The default value is 500.
8423 @item reorder-blocks-duplicate
8424 @itemx reorder-blocks-duplicate-feedback
8426 Used by basic block reordering pass to decide whether to use unconditional
8427 branch or duplicate the code on its destination. Code is duplicated when its
8428 estimated size is smaller than this value multiplied by the estimated size of
8429 unconditional jump in the hot spots of the program.
8431 The @option{reorder-block-duplicate-feedback} is used only when profile
8432 feedback is available and may be set to higher values than
8433 @option{reorder-block-duplicate} since information about the hot spots is more
8436 @item max-sched-ready-insns
8437 The maximum number of instructions ready to be issued the scheduler should
8438 consider at any given time during the first scheduling pass. Increasing
8439 values mean more thorough searches, making the compilation time increase
8440 with probably little benefit. The default value is 100.
8442 @item max-sched-region-blocks
8443 The maximum number of blocks in a region to be considered for
8444 interblock scheduling. The default value is 10.
8446 @item max-pipeline-region-blocks
8447 The maximum number of blocks in a region to be considered for
8448 pipelining in the selective scheduler. The default value is 15.
8450 @item max-sched-region-insns
8451 The maximum number of insns in a region to be considered for
8452 interblock scheduling. The default value is 100.
8454 @item max-pipeline-region-insns
8455 The maximum number of insns in a region to be considered for
8456 pipelining in the selective scheduler. The default value is 200.
8459 The minimum probability (in percents) of reaching a source block
8460 for interblock speculative scheduling. The default value is 40.
8462 @item max-sched-extend-regions-iters
8463 The maximum number of iterations through CFG to extend regions.
8464 0 - disable region extension,
8465 N - do at most N iterations.
8466 The default value is 0.
8468 @item max-sched-insn-conflict-delay
8469 The maximum conflict delay for an insn to be considered for speculative motion.
8470 The default value is 3.
8472 @item sched-spec-prob-cutoff
8473 The minimal probability of speculation success (in percents), so that
8474 speculative insn will be scheduled.
8475 The default value is 40.
8477 @item sched-mem-true-dep-cost
8478 Minimal distance (in CPU cycles) between store and load targeting same
8479 memory locations. The default value is 1.
8481 @item selsched-max-lookahead
8482 The maximum size of the lookahead window of selective scheduling. It is a
8483 depth of search for available instructions.
8484 The default value is 50.
8486 @item selsched-max-sched-times
8487 The maximum number of times that an instruction will be scheduled during
8488 selective scheduling. This is the limit on the number of iterations
8489 through which the instruction may be pipelined. The default value is 2.
8491 @item selsched-max-insns-to-rename
8492 The maximum number of best instructions in the ready list that are considered
8493 for renaming in the selective scheduler. The default value is 2.
8495 @item max-last-value-rtl
8496 The maximum size measured as number of RTLs that can be recorded in an expression
8497 in combiner for a pseudo register as last known value of that register. The default
8500 @item integer-share-limit
8501 Small integer constants can use a shared data structure, reducing the
8502 compiler's memory usage and increasing its speed. This sets the maximum
8503 value of a shared integer constant. The default value is 256.
8505 @item min-virtual-mappings
8506 Specifies the minimum number of virtual mappings in the incremental
8507 SSA updater that should be registered to trigger the virtual mappings
8508 heuristic defined by virtual-mappings-ratio. The default value is
8511 @item virtual-mappings-ratio
8512 If the number of virtual mappings is virtual-mappings-ratio bigger
8513 than the number of virtual symbols to be updated, then the incremental
8514 SSA updater switches to a full update for those symbols. The default
8517 @item ssp-buffer-size
8518 The minimum size of buffers (i.e.@: arrays) that will receive stack smashing
8519 protection when @option{-fstack-protection} is used.
8521 @item max-jump-thread-duplication-stmts
8522 Maximum number of statements allowed in a block that needs to be
8523 duplicated when threading jumps.
8525 @item max-fields-for-field-sensitive
8526 Maximum number of fields in a structure we will treat in
8527 a field sensitive manner during pointer analysis. The default is zero
8528 for -O0, and -O1 and 100 for -Os, -O2, and -O3.
8530 @item prefetch-latency
8531 Estimate on average number of instructions that are executed before
8532 prefetch finishes. The distance we prefetch ahead is proportional
8533 to this constant. Increasing this number may also lead to less
8534 streams being prefetched (see @option{simultaneous-prefetches}).
8536 @item simultaneous-prefetches
8537 Maximum number of prefetches that can run at the same time.
8539 @item l1-cache-line-size
8540 The size of cache line in L1 cache, in bytes.
8543 The size of L1 cache, in kilobytes.
8546 The size of L2 cache, in kilobytes.
8548 @item min-insn-to-prefetch-ratio
8549 The minimum ratio between the number of instructions and the
8550 number of prefetches to enable prefetching in a loop.
8552 @item prefetch-min-insn-to-mem-ratio
8553 The minimum ratio between the number of instructions and the
8554 number of memory references to enable prefetching in a loop.
8556 @item use-canonical-types
8557 Whether the compiler should use the ``canonical'' type system. By
8558 default, this should always be 1, which uses a more efficient internal
8559 mechanism for comparing types in C++ and Objective-C++. However, if
8560 bugs in the canonical type system are causing compilation failures,
8561 set this value to 0 to disable canonical types.
8563 @item switch-conversion-max-branch-ratio
8564 Switch initialization conversion will refuse to create arrays that are
8565 bigger than @option{switch-conversion-max-branch-ratio} times the number of
8566 branches in the switch.
8568 @item max-partial-antic-length
8569 Maximum length of the partial antic set computed during the tree
8570 partial redundancy elimination optimization (@option{-ftree-pre}) when
8571 optimizing at @option{-O3} and above. For some sorts of source code
8572 the enhanced partial redundancy elimination optimization can run away,
8573 consuming all of the memory available on the host machine. This
8574 parameter sets a limit on the length of the sets that are computed,
8575 which prevents the runaway behavior. Setting a value of 0 for
8576 this parameter will allow an unlimited set length.
8578 @item sccvn-max-scc-size
8579 Maximum size of a strongly connected component (SCC) during SCCVN
8580 processing. If this limit is hit, SCCVN processing for the whole
8581 function will not be done and optimizations depending on it will
8582 be disabled. The default maximum SCC size is 10000.
8584 @item ira-max-loops-num
8585 IRA uses a regional register allocation by default. If a function
8586 contains loops more than number given by the parameter, only at most
8587 given number of the most frequently executed loops will form regions
8588 for the regional register allocation. The default value of the
8591 @item ira-max-conflict-table-size
8592 Although IRA uses a sophisticated algorithm of compression conflict
8593 table, the table can be still big for huge functions. If the conflict
8594 table for a function could be more than size in MB given by the
8595 parameter, the conflict table is not built and faster, simpler, and
8596 lower quality register allocation algorithm will be used. The
8597 algorithm do not use pseudo-register conflicts. The default value of
8598 the parameter is 2000.
8600 @item ira-loop-reserved-regs
8601 IRA can be used to evaluate more accurate register pressure in loops
8602 for decision to move loop invariants (see @option{-O3}). The number
8603 of available registers reserved for some other purposes is described
8604 by this parameter. The default value of the parameter is 2 which is
8605 minimal number of registers needed for execution of typical
8606 instruction. This value is the best found from numerous experiments.
8608 @item loop-invariant-max-bbs-in-loop
8609 Loop invariant motion can be very expensive, both in compile time and
8610 in amount of needed compile time memory, with very large loops. Loops
8611 with more basic blocks than this parameter won't have loop invariant
8612 motion optimization performed on them. The default value of the
8613 parameter is 1000 for -O1 and 10000 for -O2 and above.
8615 @item max-vartrack-size
8616 Sets a maximum number of hash table slots to use during variable
8617 tracking dataflow analysis of any function. If this limit is exceeded
8618 with variable tracking at assignments enabled, analysis for that
8619 function is retried without it, after removing all debug insns from
8620 the function. If the limit is exceeded even without debug insns, var
8621 tracking analysis is completely disabled for the function. Setting
8622 the parameter to zero makes it unlimited.
8624 @item min-nondebug-insn-uid
8625 Use uids starting at this parameter for nondebug insns. The range below
8626 the parameter is reserved exclusively for debug insns created by
8627 @option{-fvar-tracking-assignments}, but debug insns may get
8628 (non-overlapping) uids above it if the reserved range is exhausted.
8630 @item ipa-sra-ptr-growth-factor
8631 IPA-SRA will replace a pointer to an aggregate with one or more new
8632 parameters only when their cumulative size is less or equal to
8633 @option{ipa-sra-ptr-growth-factor} times the size of the original
8636 @item graphite-max-nb-scop-params
8637 To avoid exponential effects in the Graphite loop transforms, the
8638 number of parameters in a Static Control Part (SCoP) is bounded. The
8639 default value is 10 parameters. A variable whose value is unknown at
8640 compile time and defined outside a SCoP is a parameter of the SCoP.
8642 @item graphite-max-bbs-per-function
8643 To avoid exponential effects in the detection of SCoPs, the size of
8644 the functions analyzed by Graphite is bounded. The default value is
8647 @item loop-block-tile-size
8648 Loop blocking or strip mining transforms, enabled with
8649 @option{-floop-block} or @option{-floop-strip-mine}, strip mine each
8650 loop in the loop nest by a given number of iterations. The strip
8651 length can be changed using the @option{loop-block-tile-size}
8652 parameter. The default value is 51 iterations.
8657 @node Preprocessor Options
8658 @section Options Controlling the Preprocessor
8659 @cindex preprocessor options
8660 @cindex options, preprocessor
8662 These options control the C preprocessor, which is run on each C source
8663 file before actual compilation.
8665 If you use the @option{-E} option, nothing is done except preprocessing.
8666 Some of these options make sense only together with @option{-E} because
8667 they cause the preprocessor output to be unsuitable for actual
8671 @item -Wp,@var{option}
8673 You can use @option{-Wp,@var{option}} to bypass the compiler driver
8674 and pass @var{option} directly through to the preprocessor. If
8675 @var{option} contains commas, it is split into multiple options at the
8676 commas. However, many options are modified, translated or interpreted
8677 by the compiler driver before being passed to the preprocessor, and
8678 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
8679 interface is undocumented and subject to change, so whenever possible
8680 you should avoid using @option{-Wp} and let the driver handle the
8683 @item -Xpreprocessor @var{option}
8684 @opindex Xpreprocessor
8685 Pass @var{option} as an option to the preprocessor. You can use this to
8686 supply system-specific preprocessor options which GCC does not know how to
8689 If you want to pass an option that takes an argument, you must use
8690 @option{-Xpreprocessor} twice, once for the option and once for the argument.
8693 @include cppopts.texi
8695 @node Assembler Options
8696 @section Passing Options to the Assembler
8698 @c prevent bad page break with this line
8699 You can pass options to the assembler.
8702 @item -Wa,@var{option}
8704 Pass @var{option} as an option to the assembler. If @var{option}
8705 contains commas, it is split into multiple options at the commas.
8707 @item -Xassembler @var{option}
8709 Pass @var{option} as an option to the assembler. You can use this to
8710 supply system-specific assembler options which GCC does not know how to
8713 If you want to pass an option that takes an argument, you must use
8714 @option{-Xassembler} twice, once for the option and once for the argument.
8719 @section Options for Linking
8720 @cindex link options
8721 @cindex options, linking
8723 These options come into play when the compiler links object files into
8724 an executable output file. They are meaningless if the compiler is
8725 not doing a link step.
8729 @item @var{object-file-name}
8730 A file name that does not end in a special recognized suffix is
8731 considered to name an object file or library. (Object files are
8732 distinguished from libraries by the linker according to the file
8733 contents.) If linking is done, these object files are used as input
8742 If any of these options is used, then the linker is not run, and
8743 object file names should not be used as arguments. @xref{Overall
8747 @item -l@var{library}
8748 @itemx -l @var{library}
8750 Search the library named @var{library} when linking. (The second
8751 alternative with the library as a separate argument is only for
8752 POSIX compliance and is not recommended.)
8754 It makes a difference where in the command you write this option; the
8755 linker searches and processes libraries and object files in the order they
8756 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
8757 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
8758 to functions in @samp{z}, those functions may not be loaded.
8760 The linker searches a standard list of directories for the library,
8761 which is actually a file named @file{lib@var{library}.a}. The linker
8762 then uses this file as if it had been specified precisely by name.
8764 The directories searched include several standard system directories
8765 plus any that you specify with @option{-L}.
8767 Normally the files found this way are library files---archive files
8768 whose members are object files. The linker handles an archive file by
8769 scanning through it for members which define symbols that have so far
8770 been referenced but not defined. But if the file that is found is an
8771 ordinary object file, it is linked in the usual fashion. The only
8772 difference between using an @option{-l} option and specifying a file name
8773 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
8774 and searches several directories.
8778 You need this special case of the @option{-l} option in order to
8779 link an Objective-C or Objective-C++ program.
8782 @opindex nostartfiles
8783 Do not use the standard system startup files when linking.
8784 The standard system libraries are used normally, unless @option{-nostdlib}
8785 or @option{-nodefaultlibs} is used.
8787 @item -nodefaultlibs
8788 @opindex nodefaultlibs
8789 Do not use the standard system libraries when linking.
8790 Only the libraries you specify will be passed to the linker, options
8791 specifying linkage of the system libraries, such as @code{-static-libgcc}
8792 or @code{-shared-libgcc}, will be ignored.
8793 The standard startup files are used normally, unless @option{-nostartfiles}
8794 is used. The compiler may generate calls to @code{memcmp},
8795 @code{memset}, @code{memcpy} and @code{memmove}.
8796 These entries are usually resolved by entries in
8797 libc. These entry points should be supplied through some other
8798 mechanism when this option is specified.
8802 Do not use the standard system startup files or libraries when linking.
8803 No startup files and only the libraries you specify will be passed to
8804 the linker, options specifying linkage of the system libraries, such as
8805 @code{-static-libgcc} or @code{-shared-libgcc}, will be ignored.
8806 The compiler may generate calls to @code{memcmp}, @code{memset},
8807 @code{memcpy} and @code{memmove}.
8808 These entries are usually resolved by entries in
8809 libc. These entry points should be supplied through some other
8810 mechanism when this option is specified.
8812 @cindex @option{-lgcc}, use with @option{-nostdlib}
8813 @cindex @option{-nostdlib} and unresolved references
8814 @cindex unresolved references and @option{-nostdlib}
8815 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
8816 @cindex @option{-nodefaultlibs} and unresolved references
8817 @cindex unresolved references and @option{-nodefaultlibs}
8818 One of the standard libraries bypassed by @option{-nostdlib} and
8819 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
8820 that GCC uses to overcome shortcomings of particular machines, or special
8821 needs for some languages.
8822 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
8823 Collection (GCC) Internals},
8824 for more discussion of @file{libgcc.a}.)
8825 In most cases, you need @file{libgcc.a} even when you want to avoid
8826 other standard libraries. In other words, when you specify @option{-nostdlib}
8827 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
8828 This ensures that you have no unresolved references to internal GCC
8829 library subroutines. (For example, @samp{__main}, used to ensure C++
8830 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
8831 GNU Compiler Collection (GCC) Internals}.)
8835 Produce a position independent executable on targets which support it.
8836 For predictable results, you must also specify the same set of options
8837 that were used to generate code (@option{-fpie}, @option{-fPIE},
8838 or model suboptions) when you specify this option.
8842 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
8843 that support it. This instructs the linker to add all symbols, not
8844 only used ones, to the dynamic symbol table. This option is needed
8845 for some uses of @code{dlopen} or to allow obtaining backtraces
8846 from within a program.
8850 Remove all symbol table and relocation information from the executable.
8854 On systems that support dynamic linking, this prevents linking with the shared
8855 libraries. On other systems, this option has no effect.
8859 Produce a shared object which can then be linked with other objects to
8860 form an executable. Not all systems support this option. For predictable
8861 results, you must also specify the same set of options that were used to
8862 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
8863 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
8864 needs to build supplementary stub code for constructors to work. On
8865 multi-libbed systems, @samp{gcc -shared} must select the correct support
8866 libraries to link against. Failing to supply the correct flags may lead
8867 to subtle defects. Supplying them in cases where they are not necessary
8870 @item -shared-libgcc
8871 @itemx -static-libgcc
8872 @opindex shared-libgcc
8873 @opindex static-libgcc
8874 On systems that provide @file{libgcc} as a shared library, these options
8875 force the use of either the shared or static version respectively.
8876 If no shared version of @file{libgcc} was built when the compiler was
8877 configured, these options have no effect.
8879 There are several situations in which an application should use the
8880 shared @file{libgcc} instead of the static version. The most common
8881 of these is when the application wishes to throw and catch exceptions
8882 across different shared libraries. In that case, each of the libraries
8883 as well as the application itself should use the shared @file{libgcc}.
8885 Therefore, the G++ and GCJ drivers automatically add
8886 @option{-shared-libgcc} whenever you build a shared library or a main
8887 executable, because C++ and Java programs typically use exceptions, so
8888 this is the right thing to do.
8890 If, instead, you use the GCC driver to create shared libraries, you may
8891 find that they will not always be linked with the shared @file{libgcc}.
8892 If GCC finds, at its configuration time, that you have a non-GNU linker
8893 or a GNU linker that does not support option @option{--eh-frame-hdr},
8894 it will link the shared version of @file{libgcc} into shared libraries
8895 by default. Otherwise, it will take advantage of the linker and optimize
8896 away the linking with the shared version of @file{libgcc}, linking with
8897 the static version of libgcc by default. This allows exceptions to
8898 propagate through such shared libraries, without incurring relocation
8899 costs at library load time.
8901 However, if a library or main executable is supposed to throw or catch
8902 exceptions, you must link it using the G++ or GCJ driver, as appropriate
8903 for the languages used in the program, or using the option
8904 @option{-shared-libgcc}, such that it is linked with the shared
8907 @item -static-libstdc++
8908 When the @command{g++} program is used to link a C++ program, it will
8909 normally automatically link against @option{libstdc++}. If
8910 @file{libstdc++} is available as a shared library, and the
8911 @option{-static} option is not used, then this will link against the
8912 shared version of @file{libstdc++}. That is normally fine. However, it
8913 is sometimes useful to freeze the version of @file{libstdc++} used by
8914 the program without going all the way to a fully static link. The
8915 @option{-static-libstdc++} option directs the @command{g++} driver to
8916 link @file{libstdc++} statically, without necessarily linking other
8917 libraries statically.
8921 Bind references to global symbols when building a shared object. Warn
8922 about any unresolved references (unless overridden by the link editor
8923 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
8926 @item -T @var{script}
8928 @cindex linker script
8929 Use @var{script} as the linker script. This option is supported by most
8930 systems using the GNU linker. On some targets, such as bare-board
8931 targets without an operating system, the @option{-T} option may be required
8932 when linking to avoid references to undefined symbols.
8934 @item -Xlinker @var{option}
8936 Pass @var{option} as an option to the linker. You can use this to
8937 supply system-specific linker options which GCC does not know how to
8940 If you want to pass an option that takes a separate argument, you must use
8941 @option{-Xlinker} twice, once for the option and once for the argument.
8942 For example, to pass @option{-assert definitions}, you must write
8943 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
8944 @option{-Xlinker "-assert definitions"}, because this passes the entire
8945 string as a single argument, which is not what the linker expects.
8947 When using the GNU linker, it is usually more convenient to pass
8948 arguments to linker options using the @option{@var{option}=@var{value}}
8949 syntax than as separate arguments. For example, you can specify
8950 @samp{-Xlinker -Map=output.map} rather than
8951 @samp{-Xlinker -Map -Xlinker output.map}. Other linkers may not support
8952 this syntax for command-line options.
8954 @item -Wl,@var{option}
8956 Pass @var{option} as an option to the linker. If @var{option} contains
8957 commas, it is split into multiple options at the commas. You can use this
8958 syntax to pass an argument to the option.
8959 For example, @samp{-Wl,-Map,output.map} passes @samp{-Map output.map} to the
8960 linker. When using the GNU linker, you can also get the same effect with
8961 @samp{-Wl,-Map=output.map}.
8963 @item -u @var{symbol}
8965 Pretend the symbol @var{symbol} is undefined, to force linking of
8966 library modules to define it. You can use @option{-u} multiple times with
8967 different symbols to force loading of additional library modules.
8970 @node Directory Options
8971 @section Options for Directory Search
8972 @cindex directory options
8973 @cindex options, directory search
8976 These options specify directories to search for header files, for
8977 libraries and for parts of the compiler:
8982 Add the directory @var{dir} to the head of the list of directories to be
8983 searched for header files. This can be used to override a system header
8984 file, substituting your own version, since these directories are
8985 searched before the system header file directories. However, you should
8986 not use this option to add directories that contain vendor-supplied
8987 system header files (use @option{-isystem} for that). If you use more than
8988 one @option{-I} option, the directories are scanned in left-to-right
8989 order; the standard system directories come after.
8991 If a standard system include directory, or a directory specified with
8992 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
8993 option will be ignored. The directory will still be searched but as a
8994 system directory at its normal position in the system include chain.
8995 This is to ensure that GCC's procedure to fix buggy system headers and
8996 the ordering for the include_next directive are not inadvertently changed.
8997 If you really need to change the search order for system directories,
8998 use the @option{-nostdinc} and/or @option{-isystem} options.
9000 @item -iplugindir=@var{dir}
9001 Set the directory to search for plugins which are passed
9002 by @option{-fplugin=@var{name}} instead of
9003 @option{-fplugin=@var{path}/@var{name}.so}. This option is not meant
9004 to be used by the user, but only passed by the driver.
9006 @item -iquote@var{dir}
9008 Add the directory @var{dir} to the head of the list of directories to
9009 be searched for header files only for the case of @samp{#include
9010 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
9011 otherwise just like @option{-I}.
9015 Add directory @var{dir} to the list of directories to be searched
9018 @item -B@var{prefix}
9020 This option specifies where to find the executables, libraries,
9021 include files, and data files of the compiler itself.
9023 The compiler driver program runs one or more of the subprograms
9024 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
9025 @var{prefix} as a prefix for each program it tries to run, both with and
9026 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
9028 For each subprogram to be run, the compiler driver first tries the
9029 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
9030 was not specified, the driver tries two standard prefixes, which are
9031 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
9032 those results in a file name that is found, the unmodified program
9033 name is searched for using the directories specified in your
9034 @env{PATH} environment variable.
9036 The compiler will check to see if the path provided by the @option{-B}
9037 refers to a directory, and if necessary it will add a directory
9038 separator character at the end of the path.
9040 @option{-B} prefixes that effectively specify directory names also apply
9041 to libraries in the linker, because the compiler translates these
9042 options into @option{-L} options for the linker. They also apply to
9043 includes files in the preprocessor, because the compiler translates these
9044 options into @option{-isystem} options for the preprocessor. In this case,
9045 the compiler appends @samp{include} to the prefix.
9047 The run-time support file @file{libgcc.a} can also be searched for using
9048 the @option{-B} prefix, if needed. If it is not found there, the two
9049 standard prefixes above are tried, and that is all. The file is left
9050 out of the link if it is not found by those means.
9052 Another way to specify a prefix much like the @option{-B} prefix is to use
9053 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
9056 As a special kludge, if the path provided by @option{-B} is
9057 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
9058 9, then it will be replaced by @file{[dir/]include}. This is to help
9059 with boot-strapping the compiler.
9061 @item -specs=@var{file}
9063 Process @var{file} after the compiler reads in the standard @file{specs}
9064 file, in order to override the defaults that the @file{gcc} driver
9065 program uses when determining what switches to pass to @file{cc1},
9066 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
9067 @option{-specs=@var{file}} can be specified on the command line, and they
9068 are processed in order, from left to right.
9070 @item --sysroot=@var{dir}
9072 Use @var{dir} as the logical root directory for headers and libraries.
9073 For example, if the compiler would normally search for headers in
9074 @file{/usr/include} and libraries in @file{/usr/lib}, it will instead
9075 search @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
9077 If you use both this option and the @option{-isysroot} option, then
9078 the @option{--sysroot} option will apply to libraries, but the
9079 @option{-isysroot} option will apply to header files.
9081 The GNU linker (beginning with version 2.16) has the necessary support
9082 for this option. If your linker does not support this option, the
9083 header file aspect of @option{--sysroot} will still work, but the
9084 library aspect will not.
9088 This option has been deprecated. Please use @option{-iquote} instead for
9089 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
9090 Any directories you specify with @option{-I} options before the @option{-I-}
9091 option are searched only for the case of @samp{#include "@var{file}"};
9092 they are not searched for @samp{#include <@var{file}>}.
9094 If additional directories are specified with @option{-I} options after
9095 the @option{-I-}, these directories are searched for all @samp{#include}
9096 directives. (Ordinarily @emph{all} @option{-I} directories are used
9099 In addition, the @option{-I-} option inhibits the use of the current
9100 directory (where the current input file came from) as the first search
9101 directory for @samp{#include "@var{file}"}. There is no way to
9102 override this effect of @option{-I-}. With @option{-I.} you can specify
9103 searching the directory which was current when the compiler was
9104 invoked. That is not exactly the same as what the preprocessor does
9105 by default, but it is often satisfactory.
9107 @option{-I-} does not inhibit the use of the standard system directories
9108 for header files. Thus, @option{-I-} and @option{-nostdinc} are
9115 @section Specifying subprocesses and the switches to pass to them
9118 @command{gcc} is a driver program. It performs its job by invoking a
9119 sequence of other programs to do the work of compiling, assembling and
9120 linking. GCC interprets its command-line parameters and uses these to
9121 deduce which programs it should invoke, and which command-line options
9122 it ought to place on their command lines. This behavior is controlled
9123 by @dfn{spec strings}. In most cases there is one spec string for each
9124 program that GCC can invoke, but a few programs have multiple spec
9125 strings to control their behavior. The spec strings built into GCC can
9126 be overridden by using the @option{-specs=} command-line switch to specify
9129 @dfn{Spec files} are plaintext files that are used to construct spec
9130 strings. They consist of a sequence of directives separated by blank
9131 lines. The type of directive is determined by the first non-whitespace
9132 character on the line and it can be one of the following:
9135 @item %@var{command}
9136 Issues a @var{command} to the spec file processor. The commands that can
9140 @item %include <@var{file}>
9142 Search for @var{file} and insert its text at the current point in the
9145 @item %include_noerr <@var{file}>
9146 @cindex %include_noerr
9147 Just like @samp{%include}, but do not generate an error message if the include
9148 file cannot be found.
9150 @item %rename @var{old_name} @var{new_name}
9152 Rename the spec string @var{old_name} to @var{new_name}.
9156 @item *[@var{spec_name}]:
9157 This tells the compiler to create, override or delete the named spec
9158 string. All lines after this directive up to the next directive or
9159 blank line are considered to be the text for the spec string. If this
9160 results in an empty string then the spec will be deleted. (Or, if the
9161 spec did not exist, then nothing will happened.) Otherwise, if the spec
9162 does not currently exist a new spec will be created. If the spec does
9163 exist then its contents will be overridden by the text of this
9164 directive, unless the first character of that text is the @samp{+}
9165 character, in which case the text will be appended to the spec.
9167 @item [@var{suffix}]:
9168 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
9169 and up to the next directive or blank line are considered to make up the
9170 spec string for the indicated suffix. When the compiler encounters an
9171 input file with the named suffix, it will processes the spec string in
9172 order to work out how to compile that file. For example:
9179 This says that any input file whose name ends in @samp{.ZZ} should be
9180 passed to the program @samp{z-compile}, which should be invoked with the
9181 command-line switch @option{-input} and with the result of performing the
9182 @samp{%i} substitution. (See below.)
9184 As an alternative to providing a spec string, the text that follows a
9185 suffix directive can be one of the following:
9188 @item @@@var{language}
9189 This says that the suffix is an alias for a known @var{language}. This is
9190 similar to using the @option{-x} command-line switch to GCC to specify a
9191 language explicitly. For example:
9198 Says that .ZZ files are, in fact, C++ source files.
9201 This causes an error messages saying:
9204 @var{name} compiler not installed on this system.
9208 GCC already has an extensive list of suffixes built into it.
9209 This directive will add an entry to the end of the list of suffixes, but
9210 since the list is searched from the end backwards, it is effectively
9211 possible to override earlier entries using this technique.
9215 GCC has the following spec strings built into it. Spec files can
9216 override these strings or create their own. Note that individual
9217 targets can also add their own spec strings to this list.
9220 asm Options to pass to the assembler
9221 asm_final Options to pass to the assembler post-processor
9222 cpp Options to pass to the C preprocessor
9223 cc1 Options to pass to the C compiler
9224 cc1plus Options to pass to the C++ compiler
9225 endfile Object files to include at the end of the link
9226 link Options to pass to the linker
9227 lib Libraries to include on the command line to the linker
9228 libgcc Decides which GCC support library to pass to the linker
9229 linker Sets the name of the linker
9230 predefines Defines to be passed to the C preprocessor
9231 signed_char Defines to pass to CPP to say whether @code{char} is signed
9233 startfile Object files to include at the start of the link
9236 Here is a small example of a spec file:
9242 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
9245 This example renames the spec called @samp{lib} to @samp{old_lib} and
9246 then overrides the previous definition of @samp{lib} with a new one.
9247 The new definition adds in some extra command-line options before
9248 including the text of the old definition.
9250 @dfn{Spec strings} are a list of command-line options to be passed to their
9251 corresponding program. In addition, the spec strings can contain
9252 @samp{%}-prefixed sequences to substitute variable text or to
9253 conditionally insert text into the command line. Using these constructs
9254 it is possible to generate quite complex command lines.
9256 Here is a table of all defined @samp{%}-sequences for spec
9257 strings. Note that spaces are not generated automatically around the
9258 results of expanding these sequences. Therefore you can concatenate them
9259 together or combine them with constant text in a single argument.
9263 Substitute one @samp{%} into the program name or argument.
9266 Substitute the name of the input file being processed.
9269 Substitute the basename of the input file being processed.
9270 This is the substring up to (and not including) the last period
9271 and not including the directory.
9274 This is the same as @samp{%b}, but include the file suffix (text after
9278 Marks the argument containing or following the @samp{%d} as a
9279 temporary file name, so that that file will be deleted if GCC exits
9280 successfully. Unlike @samp{%g}, this contributes no text to the
9283 @item %g@var{suffix}
9284 Substitute a file name that has suffix @var{suffix} and is chosen
9285 once per compilation, and mark the argument in the same way as
9286 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
9287 name is now chosen in a way that is hard to predict even when previously
9288 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
9289 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
9290 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
9291 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
9292 was simply substituted with a file name chosen once per compilation,
9293 without regard to any appended suffix (which was therefore treated
9294 just like ordinary text), making such attacks more likely to succeed.
9296 @item %u@var{suffix}
9297 Like @samp{%g}, but generates a new temporary file name even if
9298 @samp{%u@var{suffix}} was already seen.
9300 @item %U@var{suffix}
9301 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
9302 new one if there is no such last file name. In the absence of any
9303 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
9304 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
9305 would involve the generation of two distinct file names, one
9306 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
9307 simply substituted with a file name chosen for the previous @samp{%u},
9308 without regard to any appended suffix.
9310 @item %j@var{suffix}
9311 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
9312 writable, and if save-temps is off; otherwise, substitute the name
9313 of a temporary file, just like @samp{%u}. This temporary file is not
9314 meant for communication between processes, but rather as a junk
9317 @item %|@var{suffix}
9318 @itemx %m@var{suffix}
9319 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
9320 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
9321 all. These are the two most common ways to instruct a program that it
9322 should read from standard input or write to standard output. If you
9323 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
9324 construct: see for example @file{f/lang-specs.h}.
9326 @item %.@var{SUFFIX}
9327 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
9328 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
9329 terminated by the next space or %.
9332 Marks the argument containing or following the @samp{%w} as the
9333 designated output file of this compilation. This puts the argument
9334 into the sequence of arguments that @samp{%o} will substitute later.
9337 Substitutes the names of all the output files, with spaces
9338 automatically placed around them. You should write spaces
9339 around the @samp{%o} as well or the results are undefined.
9340 @samp{%o} is for use in the specs for running the linker.
9341 Input files whose names have no recognized suffix are not compiled
9342 at all, but they are included among the output files, so they will
9346 Substitutes the suffix for object files. Note that this is
9347 handled specially when it immediately follows @samp{%g, %u, or %U},
9348 because of the need for those to form complete file names. The
9349 handling is such that @samp{%O} is treated exactly as if it had already
9350 been substituted, except that @samp{%g, %u, and %U} do not currently
9351 support additional @var{suffix} characters following @samp{%O} as they would
9352 following, for example, @samp{.o}.
9355 Substitutes the standard macro predefinitions for the
9356 current target machine. Use this when running @code{cpp}.
9359 Like @samp{%p}, but puts @samp{__} before and after the name of each
9360 predefined macro, except for macros that start with @samp{__} or with
9361 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
9365 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
9366 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
9367 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
9368 and @option{-imultilib} as necessary.
9371 Current argument is the name of a library or startup file of some sort.
9372 Search for that file in a standard list of directories and substitute
9373 the full name found. The current working directory is included in the
9374 list of directories scanned.
9377 Current argument is the name of a linker script. Search for that file
9378 in the current list of directories to scan for libraries. If the file
9379 is located insert a @option{--script} option into the command line
9380 followed by the full path name found. If the file is not found then
9381 generate an error message. Note: the current working directory is not
9385 Print @var{str} as an error message. @var{str} is terminated by a newline.
9386 Use this when inconsistent options are detected.
9389 Substitute the contents of spec string @var{name} at this point.
9392 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
9394 @item %x@{@var{option}@}
9395 Accumulate an option for @samp{%X}.
9398 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
9402 Output the accumulated assembler options specified by @option{-Wa}.
9405 Output the accumulated preprocessor options specified by @option{-Wp}.
9408 Process the @code{asm} spec. This is used to compute the
9409 switches to be passed to the assembler.
9412 Process the @code{asm_final} spec. This is a spec string for
9413 passing switches to an assembler post-processor, if such a program is
9417 Process the @code{link} spec. This is the spec for computing the
9418 command line passed to the linker. Typically it will make use of the
9419 @samp{%L %G %S %D and %E} sequences.
9422 Dump out a @option{-L} option for each directory that GCC believes might
9423 contain startup files. If the target supports multilibs then the
9424 current multilib directory will be prepended to each of these paths.
9427 Process the @code{lib} spec. This is a spec string for deciding which
9428 libraries should be included on the command line to the linker.
9431 Process the @code{libgcc} spec. This is a spec string for deciding
9432 which GCC support library should be included on the command line to the linker.
9435 Process the @code{startfile} spec. This is a spec for deciding which
9436 object files should be the first ones passed to the linker. Typically
9437 this might be a file named @file{crt0.o}.
9440 Process the @code{endfile} spec. This is a spec string that specifies
9441 the last object files that will be passed to the linker.
9444 Process the @code{cpp} spec. This is used to construct the arguments
9445 to be passed to the C preprocessor.
9448 Process the @code{cc1} spec. This is used to construct the options to be
9449 passed to the actual C compiler (@samp{cc1}).
9452 Process the @code{cc1plus} spec. This is used to construct the options to be
9453 passed to the actual C++ compiler (@samp{cc1plus}).
9456 Substitute the variable part of a matched option. See below.
9457 Note that each comma in the substituted string is replaced by
9461 Remove all occurrences of @code{-S} from the command line. Note---this
9462 command is position dependent. @samp{%} commands in the spec string
9463 before this one will see @code{-S}, @samp{%} commands in the spec string
9464 after this one will not.
9466 @item %:@var{function}(@var{args})
9467 Call the named function @var{function}, passing it @var{args}.
9468 @var{args} is first processed as a nested spec string, then split
9469 into an argument vector in the usual fashion. The function returns
9470 a string which is processed as if it had appeared literally as part
9471 of the current spec.
9473 The following built-in spec functions are provided:
9477 The @code{getenv} spec function takes two arguments: an environment
9478 variable name and a string. If the environment variable is not
9479 defined, a fatal error is issued. Otherwise, the return value is the
9480 value of the environment variable concatenated with the string. For
9481 example, if @env{TOPDIR} is defined as @file{/path/to/top}, then:
9484 %:getenv(TOPDIR /include)
9487 expands to @file{/path/to/top/include}.
9489 @item @code{if-exists}
9490 The @code{if-exists} spec function takes one argument, an absolute
9491 pathname to a file. If the file exists, @code{if-exists} returns the
9492 pathname. Here is a small example of its usage:
9496 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
9499 @item @code{if-exists-else}
9500 The @code{if-exists-else} spec function is similar to the @code{if-exists}
9501 spec function, except that it takes two arguments. The first argument is
9502 an absolute pathname to a file. If the file exists, @code{if-exists-else}
9503 returns the pathname. If it does not exist, it returns the second argument.
9504 This way, @code{if-exists-else} can be used to select one file or another,
9505 based on the existence of the first. Here is a small example of its usage:
9509 crt0%O%s %:if-exists(crti%O%s) \
9510 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
9513 @item @code{replace-outfile}
9514 The @code{replace-outfile} spec function takes two arguments. It looks for the
9515 first argument in the outfiles array and replaces it with the second argument. Here
9516 is a small example of its usage:
9519 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
9522 @item @code{print-asm-header}
9523 The @code{print-asm-header} function takes no arguments and simply
9524 prints a banner like:
9530 Use "-Wa,OPTION" to pass "OPTION" to the assembler.
9533 It is used to separate compiler options from assembler options
9534 in the @option{--target-help} output.
9538 Substitutes the @code{-S} switch, if that switch was given to GCC@.
9539 If that switch was not specified, this substitutes nothing. Note that
9540 the leading dash is omitted when specifying this option, and it is
9541 automatically inserted if the substitution is performed. Thus the spec
9542 string @samp{%@{foo@}} would match the command-line option @option{-foo}
9543 and would output the command line option @option{-foo}.
9545 @item %W@{@code{S}@}
9546 Like %@{@code{S}@} but mark last argument supplied within as a file to be
9549 @item %@{@code{S}*@}
9550 Substitutes all the switches specified to GCC whose names start
9551 with @code{-S}, but which also take an argument. This is used for
9552 switches like @option{-o}, @option{-D}, @option{-I}, etc.
9553 GCC considers @option{-o foo} as being
9554 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
9555 text, including the space. Thus two arguments would be generated.
9557 @item %@{@code{S}*&@code{T}*@}
9558 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
9559 (the order of @code{S} and @code{T} in the spec is not significant).
9560 There can be any number of ampersand-separated variables; for each the
9561 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
9563 @item %@{@code{S}:@code{X}@}
9564 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
9566 @item %@{!@code{S}:@code{X}@}
9567 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
9569 @item %@{@code{S}*:@code{X}@}
9570 Substitutes @code{X} if one or more switches whose names start with
9571 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
9572 once, no matter how many such switches appeared. However, if @code{%*}
9573 appears somewhere in @code{X}, then @code{X} will be substituted once
9574 for each matching switch, with the @code{%*} replaced by the part of
9575 that switch that matched the @code{*}.
9577 @item %@{.@code{S}:@code{X}@}
9578 Substitutes @code{X}, if processing a file with suffix @code{S}.
9580 @item %@{!.@code{S}:@code{X}@}
9581 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
9583 @item %@{,@code{S}:@code{X}@}
9584 Substitutes @code{X}, if processing a file for language @code{S}.
9586 @item %@{!,@code{S}:@code{X}@}
9587 Substitutes @code{X}, if not processing a file for language @code{S}.
9589 @item %@{@code{S}|@code{P}:@code{X}@}
9590 Substitutes @code{X} if either @code{-S} or @code{-P} was given to
9591 GCC@. This may be combined with @samp{!}, @samp{.}, @samp{,}, and
9592 @code{*} sequences as well, although they have a stronger binding than
9593 the @samp{|}. If @code{%*} appears in @code{X}, all of the
9594 alternatives must be starred, and only the first matching alternative
9597 For example, a spec string like this:
9600 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
9603 will output the following command-line options from the following input
9604 command-line options:
9609 -d fred.c -foo -baz -boggle
9610 -d jim.d -bar -baz -boggle
9613 @item %@{S:X; T:Y; :D@}
9615 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
9616 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
9617 be as many clauses as you need. This may be combined with @code{.},
9618 @code{,}, @code{!}, @code{|}, and @code{*} as needed.
9623 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
9624 construct may contain other nested @samp{%} constructs or spaces, or
9625 even newlines. They are processed as usual, as described above.
9626 Trailing white space in @code{X} is ignored. White space may also
9627 appear anywhere on the left side of the colon in these constructs,
9628 except between @code{.} or @code{*} and the corresponding word.
9630 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
9631 handled specifically in these constructs. If another value of
9632 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
9633 @option{-W} switch is found later in the command line, the earlier
9634 switch value is ignored, except with @{@code{S}*@} where @code{S} is
9635 just one letter, which passes all matching options.
9637 The character @samp{|} at the beginning of the predicate text is used to
9638 indicate that a command should be piped to the following command, but
9639 only if @option{-pipe} is specified.
9641 It is built into GCC which switches take arguments and which do not.
9642 (You might think it would be useful to generalize this to allow each
9643 compiler's spec to say which switches take arguments. But this cannot
9644 be done in a consistent fashion. GCC cannot even decide which input
9645 files have been specified without knowing which switches take arguments,
9646 and it must know which input files to compile in order to tell which
9649 GCC also knows implicitly that arguments starting in @option{-l} are to be
9650 treated as compiler output files, and passed to the linker in their
9651 proper position among the other output files.
9653 @c man begin OPTIONS
9655 @node Target Options
9656 @section Specifying Target Machine and Compiler Version
9657 @cindex target options
9658 @cindex cross compiling
9659 @cindex specifying machine version
9660 @cindex specifying compiler version and target machine
9661 @cindex compiler version, specifying
9662 @cindex target machine, specifying
9664 The usual way to run GCC is to run the executable called @file{gcc}, or
9665 @file{<machine>-gcc} when cross-compiling, or
9666 @file{<machine>-gcc-<version>} to run a version other than the one that
9669 @node Submodel Options
9670 @section Hardware Models and Configurations
9671 @cindex submodel options
9672 @cindex specifying hardware config
9673 @cindex hardware models and configurations, specifying
9674 @cindex machine dependent options
9676 Each target machine types can have its own
9677 special options, starting with @samp{-m}, to choose among various
9678 hardware models or configurations---for example, 68010 vs 68020,
9679 floating coprocessor or none. A single installed version of the
9680 compiler can compile for any model or configuration, according to the
9683 Some configurations of the compiler also support additional special
9684 options, usually for compatibility with other compilers on the same
9687 @c This list is ordered alphanumerically by subsection name.
9688 @c It should be the same order and spelling as these options are listed
9689 @c in Machine Dependent Options
9695 * Blackfin Options::
9699 * DEC Alpha Options::
9700 * DEC Alpha/VMS Options::
9703 * GNU/Linux Options::
9706 * i386 and x86-64 Options::
9707 * i386 and x86-64 Windows Options::
9709 * IA-64/VMS Options::
9721 * picoChip Options::
9723 * RS/6000 and PowerPC Options::
9725 * S/390 and zSeries Options::
9730 * System V Options::
9735 * Xstormy16 Options::
9741 @subsection ARC Options
9744 These options are defined for ARC implementations:
9749 Compile code for little endian mode. This is the default.
9753 Compile code for big endian mode.
9756 @opindex mmangle-cpu
9757 Prepend the name of the cpu to all public symbol names.
9758 In multiple-processor systems, there are many ARC variants with different
9759 instruction and register set characteristics. This flag prevents code
9760 compiled for one cpu to be linked with code compiled for another.
9761 No facility exists for handling variants that are ``almost identical''.
9762 This is an all or nothing option.
9764 @item -mcpu=@var{cpu}
9766 Compile code for ARC variant @var{cpu}.
9767 Which variants are supported depend on the configuration.
9768 All variants support @option{-mcpu=base}, this is the default.
9770 @item -mtext=@var{text-section}
9771 @itemx -mdata=@var{data-section}
9772 @itemx -mrodata=@var{readonly-data-section}
9776 Put functions, data, and readonly data in @var{text-section},
9777 @var{data-section}, and @var{readonly-data-section} respectively
9778 by default. This can be overridden with the @code{section} attribute.
9779 @xref{Variable Attributes}.
9784 @subsection ARM Options
9787 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
9791 @item -mabi=@var{name}
9793 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
9794 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
9797 @opindex mapcs-frame
9798 Generate a stack frame that is compliant with the ARM Procedure Call
9799 Standard for all functions, even if this is not strictly necessary for
9800 correct execution of the code. Specifying @option{-fomit-frame-pointer}
9801 with this option will cause the stack frames not to be generated for
9802 leaf functions. The default is @option{-mno-apcs-frame}.
9806 This is a synonym for @option{-mapcs-frame}.
9809 @c not currently implemented
9810 @item -mapcs-stack-check
9811 @opindex mapcs-stack-check
9812 Generate code to check the amount of stack space available upon entry to
9813 every function (that actually uses some stack space). If there is
9814 insufficient space available then either the function
9815 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
9816 called, depending upon the amount of stack space required. The run time
9817 system is required to provide these functions. The default is
9818 @option{-mno-apcs-stack-check}, since this produces smaller code.
9820 @c not currently implemented
9822 @opindex mapcs-float
9823 Pass floating point arguments using the float point registers. This is
9824 one of the variants of the APCS@. This option is recommended if the
9825 target hardware has a floating point unit or if a lot of floating point
9826 arithmetic is going to be performed by the code. The default is
9827 @option{-mno-apcs-float}, since integer only code is slightly increased in
9828 size if @option{-mapcs-float} is used.
9830 @c not currently implemented
9831 @item -mapcs-reentrant
9832 @opindex mapcs-reentrant
9833 Generate reentrant, position independent code. The default is
9834 @option{-mno-apcs-reentrant}.
9837 @item -mthumb-interwork
9838 @opindex mthumb-interwork
9839 Generate code which supports calling between the ARM and Thumb
9840 instruction sets. Without this option the two instruction sets cannot
9841 be reliably used inside one program. The default is
9842 @option{-mno-thumb-interwork}, since slightly larger code is generated
9843 when @option{-mthumb-interwork} is specified.
9845 @item -mno-sched-prolog
9846 @opindex mno-sched-prolog
9847 Prevent the reordering of instructions in the function prolog, or the
9848 merging of those instruction with the instructions in the function's
9849 body. This means that all functions will start with a recognizable set
9850 of instructions (or in fact one of a choice from a small set of
9851 different function prologues), and this information can be used to
9852 locate the start if functions inside an executable piece of code. The
9853 default is @option{-msched-prolog}.
9855 @item -mfloat-abi=@var{name}
9857 Specifies which floating-point ABI to use. Permissible values
9858 are: @samp{soft}, @samp{softfp} and @samp{hard}.
9860 Specifying @samp{soft} causes GCC to generate output containing
9861 library calls for floating-point operations.
9862 @samp{softfp} allows the generation of code using hardware floating-point
9863 instructions, but still uses the soft-float calling conventions.
9864 @samp{hard} allows generation of floating-point instructions
9865 and uses FPU-specific calling conventions.
9867 The default depends on the specific target configuration. Note that
9868 the hard-float and soft-float ABIs are not link-compatible; you must
9869 compile your entire program with the same ABI, and link with a
9870 compatible set of libraries.
9873 @opindex mhard-float
9874 Equivalent to @option{-mfloat-abi=hard}.
9877 @opindex msoft-float
9878 Equivalent to @option{-mfloat-abi=soft}.
9880 @item -mlittle-endian
9881 @opindex mlittle-endian
9882 Generate code for a processor running in little-endian mode. This is
9883 the default for all standard configurations.
9886 @opindex mbig-endian
9887 Generate code for a processor running in big-endian mode; the default is
9888 to compile code for a little-endian processor.
9890 @item -mwords-little-endian
9891 @opindex mwords-little-endian
9892 This option only applies when generating code for big-endian processors.
9893 Generate code for a little-endian word order but a big-endian byte
9894 order. That is, a byte order of the form @samp{32107654}. Note: this
9895 option should only be used if you require compatibility with code for
9896 big-endian ARM processors generated by versions of the compiler prior to
9899 @item -mcpu=@var{name}
9901 This specifies the name of the target ARM processor. GCC uses this name
9902 to determine what kind of instructions it can emit when generating
9903 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
9904 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
9905 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
9906 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
9907 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
9909 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
9910 @samp{arm710t}, @samp{arm720t}, @samp{arm740t},
9911 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
9912 @samp{strongarm1110},
9913 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
9914 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
9915 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
9916 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
9917 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
9918 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
9919 @samp{arm1156t2-s}, @samp{arm1156t2f-s}, @samp{arm1176jz-s}, @samp{arm1176jzf-s},
9920 @samp{cortex-a5}, @samp{cortex-a8}, @samp{cortex-a9},
9921 @samp{cortex-r4}, @samp{cortex-r4f}, @samp{cortex-m4}, @samp{cortex-m3},
9924 @samp{xscale}, @samp{iwmmxt}, @samp{iwmmxt2}, @samp{ep9312}.
9926 @item -mtune=@var{name}
9928 This option is very similar to the @option{-mcpu=} option, except that
9929 instead of specifying the actual target processor type, and hence
9930 restricting which instructions can be used, it specifies that GCC should
9931 tune the performance of the code as if the target were of the type
9932 specified in this option, but still choosing the instructions that it
9933 will generate based on the cpu specified by a @option{-mcpu=} option.
9934 For some ARM implementations better performance can be obtained by using
9937 @item -march=@var{name}
9939 This specifies the name of the target ARM architecture. GCC uses this
9940 name to determine what kind of instructions it can emit when generating
9941 assembly code. This option can be used in conjunction with or instead
9942 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
9943 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
9944 @samp{armv5}, @samp{armv5t}, @samp{armv5e}, @samp{armv5te},
9945 @samp{armv6}, @samp{armv6j},
9946 @samp{armv6t2}, @samp{armv6z}, @samp{armv6zk}, @samp{armv6-m},
9947 @samp{armv7}, @samp{armv7-a}, @samp{armv7-r}, @samp{armv7-m},
9948 @samp{iwmmxt}, @samp{iwmmxt2}, @samp{ep9312}.
9950 @item -mfpu=@var{name}
9951 @itemx -mfpe=@var{number}
9952 @itemx -mfp=@var{number}
9956 This specifies what floating point hardware (or hardware emulation) is
9957 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
9958 @samp{fpe3}, @samp{maverick}, @samp{vfp}, @samp{vfpv3}, @samp{vfpv3-fp16},
9959 @samp{vfpv3-d16}, @samp{vfpv3-d16-fp16}, @samp{vfpv3xd}, @samp{vfpv3xd-fp16},
9960 @samp{neon}, @samp{neon-fp16}, @samp{vfpv4}, @samp{vfpv4-d16},
9961 @samp{fpv4-sp-d16} and @samp{neon-vfpv4}.
9962 @option{-mfp} and @option{-mfpe} are synonyms for
9963 @option{-mfpu}=@samp{fpe}@var{number}, for compatibility with older versions
9966 If @option{-msoft-float} is specified this specifies the format of
9967 floating point values.
9969 @item -mfp16-format=@var{name}
9970 @opindex mfp16-format
9971 Specify the format of the @code{__fp16} half-precision floating-point type.
9972 Permissible names are @samp{none}, @samp{ieee}, and @samp{alternative};
9973 the default is @samp{none}, in which case the @code{__fp16} type is not
9974 defined. @xref{Half-Precision}, for more information.
9976 @item -mstructure-size-boundary=@var{n}
9977 @opindex mstructure-size-boundary
9978 The size of all structures and unions will be rounded up to a multiple
9979 of the number of bits set by this option. Permissible values are 8, 32
9980 and 64. The default value varies for different toolchains. For the COFF
9981 targeted toolchain the default value is 8. A value of 64 is only allowed
9982 if the underlying ABI supports it.
9984 Specifying the larger number can produce faster, more efficient code, but
9985 can also increase the size of the program. Different values are potentially
9986 incompatible. Code compiled with one value cannot necessarily expect to
9987 work with code or libraries compiled with another value, if they exchange
9988 information using structures or unions.
9990 @item -mabort-on-noreturn
9991 @opindex mabort-on-noreturn
9992 Generate a call to the function @code{abort} at the end of a
9993 @code{noreturn} function. It will be executed if the function tries to
9997 @itemx -mno-long-calls
9998 @opindex mlong-calls
9999 @opindex mno-long-calls
10000 Tells the compiler to perform function calls by first loading the
10001 address of the function into a register and then performing a subroutine
10002 call on this register. This switch is needed if the target function
10003 will lie outside of the 64 megabyte addressing range of the offset based
10004 version of subroutine call instruction.
10006 Even if this switch is enabled, not all function calls will be turned
10007 into long calls. The heuristic is that static functions, functions
10008 which have the @samp{short-call} attribute, functions that are inside
10009 the scope of a @samp{#pragma no_long_calls} directive and functions whose
10010 definitions have already been compiled within the current compilation
10011 unit, will not be turned into long calls. The exception to this rule is
10012 that weak function definitions, functions with the @samp{long-call}
10013 attribute or the @samp{section} attribute, and functions that are within
10014 the scope of a @samp{#pragma long_calls} directive, will always be
10015 turned into long calls.
10017 This feature is not enabled by default. Specifying
10018 @option{-mno-long-calls} will restore the default behavior, as will
10019 placing the function calls within the scope of a @samp{#pragma
10020 long_calls_off} directive. Note these switches have no effect on how
10021 the compiler generates code to handle function calls via function
10024 @item -msingle-pic-base
10025 @opindex msingle-pic-base
10026 Treat the register used for PIC addressing as read-only, rather than
10027 loading it in the prologue for each function. The run-time system is
10028 responsible for initializing this register with an appropriate value
10029 before execution begins.
10031 @item -mpic-register=@var{reg}
10032 @opindex mpic-register
10033 Specify the register to be used for PIC addressing. The default is R10
10034 unless stack-checking is enabled, when R9 is used.
10036 @item -mcirrus-fix-invalid-insns
10037 @opindex mcirrus-fix-invalid-insns
10038 @opindex mno-cirrus-fix-invalid-insns
10039 Insert NOPs into the instruction stream to in order to work around
10040 problems with invalid Maverick instruction combinations. This option
10041 is only valid if the @option{-mcpu=ep9312} option has been used to
10042 enable generation of instructions for the Cirrus Maverick floating
10043 point co-processor. This option is not enabled by default, since the
10044 problem is only present in older Maverick implementations. The default
10045 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
10048 @item -mpoke-function-name
10049 @opindex mpoke-function-name
10050 Write the name of each function into the text section, directly
10051 preceding the function prologue. The generated code is similar to this:
10055 .ascii "arm_poke_function_name", 0
10058 .word 0xff000000 + (t1 - t0)
10059 arm_poke_function_name
10061 stmfd sp!, @{fp, ip, lr, pc@}
10065 When performing a stack backtrace, code can inspect the value of
10066 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
10067 location @code{pc - 12} and the top 8 bits are set, then we know that
10068 there is a function name embedded immediately preceding this location
10069 and has length @code{((pc[-3]) & 0xff000000)}.
10073 Generate code for the Thumb instruction set. The default is to
10074 use the 32-bit ARM instruction set.
10075 This option automatically enables either 16-bit Thumb-1 or
10076 mixed 16/32-bit Thumb-2 instructions based on the @option{-mcpu=@var{name}}
10077 and @option{-march=@var{name}} options. This option is not passed to the
10078 assembler. If you want to force assembler files to be interpreted as Thumb code,
10079 either add a @samp{.thumb} directive to the source or pass the @option{-mthumb}
10080 option directly to the assembler by prefixing it with @option{-Wa}.
10083 @opindex mtpcs-frame
10084 Generate a stack frame that is compliant with the Thumb Procedure Call
10085 Standard for all non-leaf functions. (A leaf function is one that does
10086 not call any other functions.) The default is @option{-mno-tpcs-frame}.
10088 @item -mtpcs-leaf-frame
10089 @opindex mtpcs-leaf-frame
10090 Generate a stack frame that is compliant with the Thumb Procedure Call
10091 Standard for all leaf functions. (A leaf function is one that does
10092 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
10094 @item -mcallee-super-interworking
10095 @opindex mcallee-super-interworking
10096 Gives all externally visible functions in the file being compiled an ARM
10097 instruction set header which switches to Thumb mode before executing the
10098 rest of the function. This allows these functions to be called from
10099 non-interworking code. This option is not valid in AAPCS configurations
10100 because interworking is enabled by default.
10102 @item -mcaller-super-interworking
10103 @opindex mcaller-super-interworking
10104 Allows calls via function pointers (including virtual functions) to
10105 execute correctly regardless of whether the target code has been
10106 compiled for interworking or not. There is a small overhead in the cost
10107 of executing a function pointer if this option is enabled. This option
10108 is not valid in AAPCS configurations because interworking is enabled
10111 @item -mtp=@var{name}
10113 Specify the access model for the thread local storage pointer. The valid
10114 models are @option{soft}, which generates calls to @code{__aeabi_read_tp},
10115 @option{cp15}, which fetches the thread pointer from @code{cp15} directly
10116 (supported in the arm6k architecture), and @option{auto}, which uses the
10117 best available method for the selected processor. The default setting is
10120 @item -mword-relocations
10121 @opindex mword-relocations
10122 Only generate absolute relocations on word sized values (i.e. R_ARM_ABS32).
10123 This is enabled by default on targets (uClinux, SymbianOS) where the runtime
10124 loader imposes this restriction, and when @option{-fpic} or @option{-fPIC}
10127 @item -mfix-cortex-m3-ldrd
10128 @opindex mfix-cortex-m3-ldrd
10129 Some Cortex-M3 cores can cause data corruption when @code{ldrd} instructions
10130 with overlapping destination and base registers are used. This option avoids
10131 generating these instructions. This option is enabled by default when
10132 @option{-mcpu=cortex-m3} is specified.
10137 @subsection AVR Options
10138 @cindex AVR Options
10140 These options are defined for AVR implementations:
10143 @item -mmcu=@var{mcu}
10145 Specify ATMEL AVR instruction set or MCU type.
10147 Instruction set avr1 is for the minimal AVR core, not supported by the C
10148 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
10149 attiny11, attiny12, attiny15, attiny28).
10151 Instruction set avr2 (default) is for the classic AVR core with up to
10152 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
10153 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
10154 at90c8534, at90s8535).
10156 Instruction set avr3 is for the classic AVR core with up to 128K program
10157 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
10159 Instruction set avr4 is for the enhanced AVR core with up to 8K program
10160 memory space (MCU types: atmega8, atmega83, atmega85).
10162 Instruction set avr5 is for the enhanced AVR core with up to 128K program
10163 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
10164 atmega64, atmega128, at43usb355, at94k).
10166 @item -mno-interrupts
10167 @opindex mno-interrupts
10168 Generated code is not compatible with hardware interrupts.
10169 Code size will be smaller.
10171 @item -mcall-prologues
10172 @opindex mcall-prologues
10173 Functions prologues/epilogues expanded as call to appropriate
10174 subroutines. Code size will be smaller.
10177 @opindex mtiny-stack
10178 Change only the low 8 bits of the stack pointer.
10182 Assume int to be 8 bit integer. This affects the sizes of all types: A
10183 char will be 1 byte, an int will be 1 byte, a long will be 2 bytes
10184 and long long will be 4 bytes. Please note that this option does not
10185 comply to the C standards, but it will provide you with smaller code
10189 @node Blackfin Options
10190 @subsection Blackfin Options
10191 @cindex Blackfin Options
10194 @item -mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]}
10196 Specifies the name of the target Blackfin processor. Currently, @var{cpu}
10197 can be one of @samp{bf512}, @samp{bf514}, @samp{bf516}, @samp{bf518},
10198 @samp{bf522}, @samp{bf523}, @samp{bf524}, @samp{bf525}, @samp{bf526},
10199 @samp{bf527}, @samp{bf531}, @samp{bf532}, @samp{bf533},
10200 @samp{bf534}, @samp{bf536}, @samp{bf537}, @samp{bf538}, @samp{bf539},
10201 @samp{bf542}, @samp{bf544}, @samp{bf547}, @samp{bf548}, @samp{bf549},
10202 @samp{bf542m}, @samp{bf544m}, @samp{bf547m}, @samp{bf548m}, @samp{bf549m},
10204 The optional @var{sirevision} specifies the silicon revision of the target
10205 Blackfin processor. Any workarounds available for the targeted silicon revision
10206 will be enabled. If @var{sirevision} is @samp{none}, no workarounds are enabled.
10207 If @var{sirevision} is @samp{any}, all workarounds for the targeted processor
10208 will be enabled. The @code{__SILICON_REVISION__} macro is defined to two
10209 hexadecimal digits representing the major and minor numbers in the silicon
10210 revision. If @var{sirevision} is @samp{none}, the @code{__SILICON_REVISION__}
10211 is not defined. If @var{sirevision} is @samp{any}, the
10212 @code{__SILICON_REVISION__} is defined to be @code{0xffff}.
10213 If this optional @var{sirevision} is not used, GCC assumes the latest known
10214 silicon revision of the targeted Blackfin processor.
10216 Support for @samp{bf561} is incomplete. For @samp{bf561},
10217 Only the processor macro is defined.
10218 Without this option, @samp{bf532} is used as the processor by default.
10219 The corresponding predefined processor macros for @var{cpu} is to
10220 be defined. And for @samp{bfin-elf} toolchain, this causes the hardware BSP
10221 provided by libgloss to be linked in if @option{-msim} is not given.
10225 Specifies that the program will be run on the simulator. This causes
10226 the simulator BSP provided by libgloss to be linked in. This option
10227 has effect only for @samp{bfin-elf} toolchain.
10228 Certain other options, such as @option{-mid-shared-library} and
10229 @option{-mfdpic}, imply @option{-msim}.
10231 @item -momit-leaf-frame-pointer
10232 @opindex momit-leaf-frame-pointer
10233 Don't keep the frame pointer in a register for leaf functions. This
10234 avoids the instructions to save, set up and restore frame pointers and
10235 makes an extra register available in leaf functions. The option
10236 @option{-fomit-frame-pointer} removes the frame pointer for all functions
10237 which might make debugging harder.
10239 @item -mspecld-anomaly
10240 @opindex mspecld-anomaly
10241 When enabled, the compiler will ensure that the generated code does not
10242 contain speculative loads after jump instructions. If this option is used,
10243 @code{__WORKAROUND_SPECULATIVE_LOADS} is defined.
10245 @item -mno-specld-anomaly
10246 @opindex mno-specld-anomaly
10247 Don't generate extra code to prevent speculative loads from occurring.
10249 @item -mcsync-anomaly
10250 @opindex mcsync-anomaly
10251 When enabled, the compiler will ensure that the generated code does not
10252 contain CSYNC or SSYNC instructions too soon after conditional branches.
10253 If this option is used, @code{__WORKAROUND_SPECULATIVE_SYNCS} is defined.
10255 @item -mno-csync-anomaly
10256 @opindex mno-csync-anomaly
10257 Don't generate extra code to prevent CSYNC or SSYNC instructions from
10258 occurring too soon after a conditional branch.
10262 When enabled, the compiler is free to take advantage of the knowledge that
10263 the entire program fits into the low 64k of memory.
10266 @opindex mno-low-64k
10267 Assume that the program is arbitrarily large. This is the default.
10269 @item -mstack-check-l1
10270 @opindex mstack-check-l1
10271 Do stack checking using information placed into L1 scratchpad memory by the
10274 @item -mid-shared-library
10275 @opindex mid-shared-library
10276 Generate code that supports shared libraries via the library ID method.
10277 This allows for execute in place and shared libraries in an environment
10278 without virtual memory management. This option implies @option{-fPIC}.
10279 With a @samp{bfin-elf} target, this option implies @option{-msim}.
10281 @item -mno-id-shared-library
10282 @opindex mno-id-shared-library
10283 Generate code that doesn't assume ID based shared libraries are being used.
10284 This is the default.
10286 @item -mleaf-id-shared-library
10287 @opindex mleaf-id-shared-library
10288 Generate code that supports shared libraries via the library ID method,
10289 but assumes that this library or executable won't link against any other
10290 ID shared libraries. That allows the compiler to use faster code for jumps
10293 @item -mno-leaf-id-shared-library
10294 @opindex mno-leaf-id-shared-library
10295 Do not assume that the code being compiled won't link against any ID shared
10296 libraries. Slower code will be generated for jump and call insns.
10298 @item -mshared-library-id=n
10299 @opindex mshared-library-id
10300 Specified the identification number of the ID based shared library being
10301 compiled. Specifying a value of 0 will generate more compact code, specifying
10302 other values will force the allocation of that number to the current
10303 library but is no more space or time efficient than omitting this option.
10307 Generate code that allows the data segment to be located in a different
10308 area of memory from the text segment. This allows for execute in place in
10309 an environment without virtual memory management by eliminating relocations
10310 against the text section.
10312 @item -mno-sep-data
10313 @opindex mno-sep-data
10314 Generate code that assumes that the data segment follows the text segment.
10315 This is the default.
10318 @itemx -mno-long-calls
10319 @opindex mlong-calls
10320 @opindex mno-long-calls
10321 Tells the compiler to perform function calls by first loading the
10322 address of the function into a register and then performing a subroutine
10323 call on this register. This switch is needed if the target function
10324 will lie outside of the 24 bit addressing range of the offset based
10325 version of subroutine call instruction.
10327 This feature is not enabled by default. Specifying
10328 @option{-mno-long-calls} will restore the default behavior. Note these
10329 switches have no effect on how the compiler generates code to handle
10330 function calls via function pointers.
10334 Link with the fast floating-point library. This library relaxes some of
10335 the IEEE floating-point standard's rules for checking inputs against
10336 Not-a-Number (NAN), in the interest of performance.
10339 @opindex minline-plt
10340 Enable inlining of PLT entries in function calls to functions that are
10341 not known to bind locally. It has no effect without @option{-mfdpic}.
10344 @opindex mmulticore
10345 Build standalone application for multicore Blackfin processor. Proper
10346 start files and link scripts will be used to support multicore.
10347 This option defines @code{__BFIN_MULTICORE}. It can only be used with
10348 @option{-mcpu=bf561@r{[}-@var{sirevision}@r{]}}. It can be used with
10349 @option{-mcorea} or @option{-mcoreb}. If it's used without
10350 @option{-mcorea} or @option{-mcoreb}, single application/dual core
10351 programming model is used. In this model, the main function of Core B
10352 should be named as coreb_main. If it's used with @option{-mcorea} or
10353 @option{-mcoreb}, one application per core programming model is used.
10354 If this option is not used, single core application programming
10359 Build standalone application for Core A of BF561 when using
10360 one application per core programming model. Proper start files
10361 and link scripts will be used to support Core A. This option
10362 defines @code{__BFIN_COREA}. It must be used with @option{-mmulticore}.
10366 Build standalone application for Core B of BF561 when using
10367 one application per core programming model. Proper start files
10368 and link scripts will be used to support Core B. This option
10369 defines @code{__BFIN_COREB}. When this option is used, coreb_main
10370 should be used instead of main. It must be used with
10371 @option{-mmulticore}.
10375 Build standalone application for SDRAM. Proper start files and
10376 link scripts will be used to put the application into SDRAM.
10377 Loader should initialize SDRAM before loading the application
10378 into SDRAM. This option defines @code{__BFIN_SDRAM}.
10382 Assume that ICPLBs are enabled at runtime. This has an effect on certain
10383 anomaly workarounds. For Linux targets, the default is to assume ICPLBs
10384 are enabled; for standalone applications the default is off.
10388 @subsection CRIS Options
10389 @cindex CRIS Options
10391 These options are defined specifically for the CRIS ports.
10394 @item -march=@var{architecture-type}
10395 @itemx -mcpu=@var{architecture-type}
10398 Generate code for the specified architecture. The choices for
10399 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
10400 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
10401 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
10404 @item -mtune=@var{architecture-type}
10406 Tune to @var{architecture-type} everything applicable about the generated
10407 code, except for the ABI and the set of available instructions. The
10408 choices for @var{architecture-type} are the same as for
10409 @option{-march=@var{architecture-type}}.
10411 @item -mmax-stack-frame=@var{n}
10412 @opindex mmax-stack-frame
10413 Warn when the stack frame of a function exceeds @var{n} bytes.
10419 The options @option{-metrax4} and @option{-metrax100} are synonyms for
10420 @option{-march=v3} and @option{-march=v8} respectively.
10422 @item -mmul-bug-workaround
10423 @itemx -mno-mul-bug-workaround
10424 @opindex mmul-bug-workaround
10425 @opindex mno-mul-bug-workaround
10426 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
10427 models where it applies. This option is active by default.
10431 Enable CRIS-specific verbose debug-related information in the assembly
10432 code. This option also has the effect to turn off the @samp{#NO_APP}
10433 formatted-code indicator to the assembler at the beginning of the
10438 Do not use condition-code results from previous instruction; always emit
10439 compare and test instructions before use of condition codes.
10441 @item -mno-side-effects
10442 @opindex mno-side-effects
10443 Do not emit instructions with side-effects in addressing modes other than
10446 @item -mstack-align
10447 @itemx -mno-stack-align
10448 @itemx -mdata-align
10449 @itemx -mno-data-align
10450 @itemx -mconst-align
10451 @itemx -mno-const-align
10452 @opindex mstack-align
10453 @opindex mno-stack-align
10454 @opindex mdata-align
10455 @opindex mno-data-align
10456 @opindex mconst-align
10457 @opindex mno-const-align
10458 These options (no-options) arranges (eliminate arrangements) for the
10459 stack-frame, individual data and constants to be aligned for the maximum
10460 single data access size for the chosen CPU model. The default is to
10461 arrange for 32-bit alignment. ABI details such as structure layout are
10462 not affected by these options.
10470 Similar to the stack- data- and const-align options above, these options
10471 arrange for stack-frame, writable data and constants to all be 32-bit,
10472 16-bit or 8-bit aligned. The default is 32-bit alignment.
10474 @item -mno-prologue-epilogue
10475 @itemx -mprologue-epilogue
10476 @opindex mno-prologue-epilogue
10477 @opindex mprologue-epilogue
10478 With @option{-mno-prologue-epilogue}, the normal function prologue and
10479 epilogue that sets up the stack-frame are omitted and no return
10480 instructions or return sequences are generated in the code. Use this
10481 option only together with visual inspection of the compiled code: no
10482 warnings or errors are generated when call-saved registers must be saved,
10483 or storage for local variable needs to be allocated.
10487 @opindex mno-gotplt
10489 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
10490 instruction sequences that load addresses for functions from the PLT part
10491 of the GOT rather than (traditional on other architectures) calls to the
10492 PLT@. The default is @option{-mgotplt}.
10496 Legacy no-op option only recognized with the cris-axis-elf and
10497 cris-axis-linux-gnu targets.
10501 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
10505 This option, recognized for the cris-axis-elf arranges
10506 to link with input-output functions from a simulator library. Code,
10507 initialized data and zero-initialized data are allocated consecutively.
10511 Like @option{-sim}, but pass linker options to locate initialized data at
10512 0x40000000 and zero-initialized data at 0x80000000.
10516 @subsection CRX Options
10517 @cindex CRX Options
10519 These options are defined specifically for the CRX ports.
10525 Enable the use of multiply-accumulate instructions. Disabled by default.
10528 @opindex mpush-args
10529 Push instructions will be used to pass outgoing arguments when functions
10530 are called. Enabled by default.
10533 @node Darwin Options
10534 @subsection Darwin Options
10535 @cindex Darwin options
10537 These options are defined for all architectures running the Darwin operating
10540 FSF GCC on Darwin does not create ``fat'' object files; it will create
10541 an object file for the single architecture that it was built to
10542 target. Apple's GCC on Darwin does create ``fat'' files if multiple
10543 @option{-arch} options are used; it does so by running the compiler or
10544 linker multiple times and joining the results together with
10547 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
10548 @samp{i686}) is determined by the flags that specify the ISA
10549 that GCC is targetting, like @option{-mcpu} or @option{-march}. The
10550 @option{-force_cpusubtype_ALL} option can be used to override this.
10552 The Darwin tools vary in their behavior when presented with an ISA
10553 mismatch. The assembler, @file{as}, will only permit instructions to
10554 be used that are valid for the subtype of the file it is generating,
10555 so you cannot put 64-bit instructions in a @samp{ppc750} object file.
10556 The linker for shared libraries, @file{/usr/bin/libtool}, will fail
10557 and print an error if asked to create a shared library with a less
10558 restrictive subtype than its input files (for instance, trying to put
10559 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
10560 for executables, @file{ld}, will quietly give the executable the most
10561 restrictive subtype of any of its input files.
10566 Add the framework directory @var{dir} to the head of the list of
10567 directories to be searched for header files. These directories are
10568 interleaved with those specified by @option{-I} options and are
10569 scanned in a left-to-right order.
10571 A framework directory is a directory with frameworks in it. A
10572 framework is a directory with a @samp{"Headers"} and/or
10573 @samp{"PrivateHeaders"} directory contained directly in it that ends
10574 in @samp{".framework"}. The name of a framework is the name of this
10575 directory excluding the @samp{".framework"}. Headers associated with
10576 the framework are found in one of those two directories, with
10577 @samp{"Headers"} being searched first. A subframework is a framework
10578 directory that is in a framework's @samp{"Frameworks"} directory.
10579 Includes of subframework headers can only appear in a header of a
10580 framework that contains the subframework, or in a sibling subframework
10581 header. Two subframeworks are siblings if they occur in the same
10582 framework. A subframework should not have the same name as a
10583 framework, a warning will be issued if this is violated. Currently a
10584 subframework cannot have subframeworks, in the future, the mechanism
10585 may be extended to support this. The standard frameworks can be found
10586 in @samp{"/System/Library/Frameworks"} and
10587 @samp{"/Library/Frameworks"}. An example include looks like
10588 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
10589 the name of the framework and header.h is found in the
10590 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
10592 @item -iframework@var{dir}
10593 @opindex iframework
10594 Like @option{-F} except the directory is a treated as a system
10595 directory. The main difference between this @option{-iframework} and
10596 @option{-F} is that with @option{-iframework} the compiler does not
10597 warn about constructs contained within header files found via
10598 @var{dir}. This option is valid only for the C family of languages.
10602 Emit debugging information for symbols that are used. For STABS
10603 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
10604 This is by default ON@.
10608 Emit debugging information for all symbols and types.
10610 @item -mmacosx-version-min=@var{version}
10611 The earliest version of MacOS X that this executable will run on
10612 is @var{version}. Typical values of @var{version} include @code{10.1},
10613 @code{10.2}, and @code{10.3.9}.
10615 If the compiler was built to use the system's headers by default,
10616 then the default for this option is the system version on which the
10617 compiler is running, otherwise the default is to make choices which
10618 are compatible with as many systems and code bases as possible.
10622 Enable kernel development mode. The @option{-mkernel} option sets
10623 @option{-static}, @option{-fno-common}, @option{-fno-cxa-atexit},
10624 @option{-fno-exceptions}, @option{-fno-non-call-exceptions},
10625 @option{-fapple-kext}, @option{-fno-weak} and @option{-fno-rtti} where
10626 applicable. This mode also sets @option{-mno-altivec},
10627 @option{-msoft-float}, @option{-fno-builtin} and
10628 @option{-mlong-branch} for PowerPC targets.
10630 @item -mone-byte-bool
10631 @opindex mone-byte-bool
10632 Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}.
10633 By default @samp{sizeof(bool)} is @samp{4} when compiling for
10634 Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this
10635 option has no effect on x86.
10637 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
10638 to generate code that is not binary compatible with code generated
10639 without that switch. Using this switch may require recompiling all
10640 other modules in a program, including system libraries. Use this
10641 switch to conform to a non-default data model.
10643 @item -mfix-and-continue
10644 @itemx -ffix-and-continue
10645 @itemx -findirect-data
10646 @opindex mfix-and-continue
10647 @opindex ffix-and-continue
10648 @opindex findirect-data
10649 Generate code suitable for fast turn around development. Needed to
10650 enable gdb to dynamically load @code{.o} files into already running
10651 programs. @option{-findirect-data} and @option{-ffix-and-continue}
10652 are provided for backwards compatibility.
10656 Loads all members of static archive libraries.
10657 See man ld(1) for more information.
10659 @item -arch_errors_fatal
10660 @opindex arch_errors_fatal
10661 Cause the errors having to do with files that have the wrong architecture
10664 @item -bind_at_load
10665 @opindex bind_at_load
10666 Causes the output file to be marked such that the dynamic linker will
10667 bind all undefined references when the file is loaded or launched.
10671 Produce a Mach-o bundle format file.
10672 See man ld(1) for more information.
10674 @item -bundle_loader @var{executable}
10675 @opindex bundle_loader
10676 This option specifies the @var{executable} that will be loading the build
10677 output file being linked. See man ld(1) for more information.
10680 @opindex dynamiclib
10681 When passed this option, GCC will produce a dynamic library instead of
10682 an executable when linking, using the Darwin @file{libtool} command.
10684 @item -force_cpusubtype_ALL
10685 @opindex force_cpusubtype_ALL
10686 This causes GCC's output file to have the @var{ALL} subtype, instead of
10687 one controlled by the @option{-mcpu} or @option{-march} option.
10689 @item -allowable_client @var{client_name}
10690 @itemx -client_name
10691 @itemx -compatibility_version
10692 @itemx -current_version
10694 @itemx -dependency-file
10696 @itemx -dylinker_install_name
10698 @itemx -exported_symbols_list
10700 @itemx -flat_namespace
10701 @itemx -force_flat_namespace
10702 @itemx -headerpad_max_install_names
10705 @itemx -install_name
10706 @itemx -keep_private_externs
10707 @itemx -multi_module
10708 @itemx -multiply_defined
10709 @itemx -multiply_defined_unused
10711 @itemx -no_dead_strip_inits_and_terms
10712 @itemx -nofixprebinding
10713 @itemx -nomultidefs
10715 @itemx -noseglinkedit
10716 @itemx -pagezero_size
10718 @itemx -prebind_all_twolevel_modules
10719 @itemx -private_bundle
10720 @itemx -read_only_relocs
10722 @itemx -sectobjectsymbols
10726 @itemx -sectobjectsymbols
10729 @itemx -segs_read_only_addr
10730 @itemx -segs_read_write_addr
10731 @itemx -seg_addr_table
10732 @itemx -seg_addr_table_filename
10733 @itemx -seglinkedit
10735 @itemx -segs_read_only_addr
10736 @itemx -segs_read_write_addr
10737 @itemx -single_module
10739 @itemx -sub_library
10740 @itemx -sub_umbrella
10741 @itemx -twolevel_namespace
10744 @itemx -unexported_symbols_list
10745 @itemx -weak_reference_mismatches
10746 @itemx -whatsloaded
10747 @opindex allowable_client
10748 @opindex client_name
10749 @opindex compatibility_version
10750 @opindex current_version
10751 @opindex dead_strip
10752 @opindex dependency-file
10753 @opindex dylib_file
10754 @opindex dylinker_install_name
10756 @opindex exported_symbols_list
10758 @opindex flat_namespace
10759 @opindex force_flat_namespace
10760 @opindex headerpad_max_install_names
10761 @opindex image_base
10763 @opindex install_name
10764 @opindex keep_private_externs
10765 @opindex multi_module
10766 @opindex multiply_defined
10767 @opindex multiply_defined_unused
10768 @opindex noall_load
10769 @opindex no_dead_strip_inits_and_terms
10770 @opindex nofixprebinding
10771 @opindex nomultidefs
10773 @opindex noseglinkedit
10774 @opindex pagezero_size
10776 @opindex prebind_all_twolevel_modules
10777 @opindex private_bundle
10778 @opindex read_only_relocs
10780 @opindex sectobjectsymbols
10783 @opindex sectcreate
10784 @opindex sectobjectsymbols
10787 @opindex segs_read_only_addr
10788 @opindex segs_read_write_addr
10789 @opindex seg_addr_table
10790 @opindex seg_addr_table_filename
10791 @opindex seglinkedit
10793 @opindex segs_read_only_addr
10794 @opindex segs_read_write_addr
10795 @opindex single_module
10797 @opindex sub_library
10798 @opindex sub_umbrella
10799 @opindex twolevel_namespace
10802 @opindex unexported_symbols_list
10803 @opindex weak_reference_mismatches
10804 @opindex whatsloaded
10805 These options are passed to the Darwin linker. The Darwin linker man page
10806 describes them in detail.
10809 @node DEC Alpha Options
10810 @subsection DEC Alpha Options
10812 These @samp{-m} options are defined for the DEC Alpha implementations:
10815 @item -mno-soft-float
10816 @itemx -msoft-float
10817 @opindex mno-soft-float
10818 @opindex msoft-float
10819 Use (do not use) the hardware floating-point instructions for
10820 floating-point operations. When @option{-msoft-float} is specified,
10821 functions in @file{libgcc.a} will be used to perform floating-point
10822 operations. Unless they are replaced by routines that emulate the
10823 floating-point operations, or compiled in such a way as to call such
10824 emulations routines, these routines will issue floating-point
10825 operations. If you are compiling for an Alpha without floating-point
10826 operations, you must ensure that the library is built so as not to call
10829 Note that Alpha implementations without floating-point operations are
10830 required to have floating-point registers.
10833 @itemx -mno-fp-regs
10835 @opindex mno-fp-regs
10836 Generate code that uses (does not use) the floating-point register set.
10837 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
10838 register set is not used, floating point operands are passed in integer
10839 registers as if they were integers and floating-point results are passed
10840 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
10841 so any function with a floating-point argument or return value called by code
10842 compiled with @option{-mno-fp-regs} must also be compiled with that
10845 A typical use of this option is building a kernel that does not use,
10846 and hence need not save and restore, any floating-point registers.
10850 The Alpha architecture implements floating-point hardware optimized for
10851 maximum performance. It is mostly compliant with the IEEE floating
10852 point standard. However, for full compliance, software assistance is
10853 required. This option generates code fully IEEE compliant code
10854 @emph{except} that the @var{inexact-flag} is not maintained (see below).
10855 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
10856 defined during compilation. The resulting code is less efficient but is
10857 able to correctly support denormalized numbers and exceptional IEEE
10858 values such as not-a-number and plus/minus infinity. Other Alpha
10859 compilers call this option @option{-ieee_with_no_inexact}.
10861 @item -mieee-with-inexact
10862 @opindex mieee-with-inexact
10863 This is like @option{-mieee} except the generated code also maintains
10864 the IEEE @var{inexact-flag}. Turning on this option causes the
10865 generated code to implement fully-compliant IEEE math. In addition to
10866 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
10867 macro. On some Alpha implementations the resulting code may execute
10868 significantly slower than the code generated by default. Since there is
10869 very little code that depends on the @var{inexact-flag}, you should
10870 normally not specify this option. Other Alpha compilers call this
10871 option @option{-ieee_with_inexact}.
10873 @item -mfp-trap-mode=@var{trap-mode}
10874 @opindex mfp-trap-mode
10875 This option controls what floating-point related traps are enabled.
10876 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
10877 The trap mode can be set to one of four values:
10881 This is the default (normal) setting. The only traps that are enabled
10882 are the ones that cannot be disabled in software (e.g., division by zero
10886 In addition to the traps enabled by @samp{n}, underflow traps are enabled
10890 Like @samp{u}, but the instructions are marked to be safe for software
10891 completion (see Alpha architecture manual for details).
10894 Like @samp{su}, but inexact traps are enabled as well.
10897 @item -mfp-rounding-mode=@var{rounding-mode}
10898 @opindex mfp-rounding-mode
10899 Selects the IEEE rounding mode. Other Alpha compilers call this option
10900 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
10905 Normal IEEE rounding mode. Floating point numbers are rounded towards
10906 the nearest machine number or towards the even machine number in case
10910 Round towards minus infinity.
10913 Chopped rounding mode. Floating point numbers are rounded towards zero.
10916 Dynamic rounding mode. A field in the floating point control register
10917 (@var{fpcr}, see Alpha architecture reference manual) controls the
10918 rounding mode in effect. The C library initializes this register for
10919 rounding towards plus infinity. Thus, unless your program modifies the
10920 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
10923 @item -mtrap-precision=@var{trap-precision}
10924 @opindex mtrap-precision
10925 In the Alpha architecture, floating point traps are imprecise. This
10926 means without software assistance it is impossible to recover from a
10927 floating trap and program execution normally needs to be terminated.
10928 GCC can generate code that can assist operating system trap handlers
10929 in determining the exact location that caused a floating point trap.
10930 Depending on the requirements of an application, different levels of
10931 precisions can be selected:
10935 Program precision. This option is the default and means a trap handler
10936 can only identify which program caused a floating point exception.
10939 Function precision. The trap handler can determine the function that
10940 caused a floating point exception.
10943 Instruction precision. The trap handler can determine the exact
10944 instruction that caused a floating point exception.
10947 Other Alpha compilers provide the equivalent options called
10948 @option{-scope_safe} and @option{-resumption_safe}.
10950 @item -mieee-conformant
10951 @opindex mieee-conformant
10952 This option marks the generated code as IEEE conformant. You must not
10953 use this option unless you also specify @option{-mtrap-precision=i} and either
10954 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
10955 is to emit the line @samp{.eflag 48} in the function prologue of the
10956 generated assembly file. Under DEC Unix, this has the effect that
10957 IEEE-conformant math library routines will be linked in.
10959 @item -mbuild-constants
10960 @opindex mbuild-constants
10961 Normally GCC examines a 32- or 64-bit integer constant to
10962 see if it can construct it from smaller constants in two or three
10963 instructions. If it cannot, it will output the constant as a literal and
10964 generate code to load it from the data segment at runtime.
10966 Use this option to require GCC to construct @emph{all} integer constants
10967 using code, even if it takes more instructions (the maximum is six).
10969 You would typically use this option to build a shared library dynamic
10970 loader. Itself a shared library, it must relocate itself in memory
10971 before it can find the variables and constants in its own data segment.
10977 Select whether to generate code to be assembled by the vendor-supplied
10978 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
10996 Indicate whether GCC should generate code to use the optional BWX,
10997 CIX, FIX and MAX instruction sets. The default is to use the instruction
10998 sets supported by the CPU type specified via @option{-mcpu=} option or that
10999 of the CPU on which GCC was built if none was specified.
11002 @itemx -mfloat-ieee
11003 @opindex mfloat-vax
11004 @opindex mfloat-ieee
11005 Generate code that uses (does not use) VAX F and G floating point
11006 arithmetic instead of IEEE single and double precision.
11008 @item -mexplicit-relocs
11009 @itemx -mno-explicit-relocs
11010 @opindex mexplicit-relocs
11011 @opindex mno-explicit-relocs
11012 Older Alpha assemblers provided no way to generate symbol relocations
11013 except via assembler macros. Use of these macros does not allow
11014 optimal instruction scheduling. GNU binutils as of version 2.12
11015 supports a new syntax that allows the compiler to explicitly mark
11016 which relocations should apply to which instructions. This option
11017 is mostly useful for debugging, as GCC detects the capabilities of
11018 the assembler when it is built and sets the default accordingly.
11021 @itemx -mlarge-data
11022 @opindex msmall-data
11023 @opindex mlarge-data
11024 When @option{-mexplicit-relocs} is in effect, static data is
11025 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
11026 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
11027 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
11028 16-bit relocations off of the @code{$gp} register. This limits the
11029 size of the small data area to 64KB, but allows the variables to be
11030 directly accessed via a single instruction.
11032 The default is @option{-mlarge-data}. With this option the data area
11033 is limited to just below 2GB@. Programs that require more than 2GB of
11034 data must use @code{malloc} or @code{mmap} to allocate the data in the
11035 heap instead of in the program's data segment.
11037 When generating code for shared libraries, @option{-fpic} implies
11038 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
11041 @itemx -mlarge-text
11042 @opindex msmall-text
11043 @opindex mlarge-text
11044 When @option{-msmall-text} is used, the compiler assumes that the
11045 code of the entire program (or shared library) fits in 4MB, and is
11046 thus reachable with a branch instruction. When @option{-msmall-data}
11047 is used, the compiler can assume that all local symbols share the
11048 same @code{$gp} value, and thus reduce the number of instructions
11049 required for a function call from 4 to 1.
11051 The default is @option{-mlarge-text}.
11053 @item -mcpu=@var{cpu_type}
11055 Set the instruction set and instruction scheduling parameters for
11056 machine type @var{cpu_type}. You can specify either the @samp{EV}
11057 style name or the corresponding chip number. GCC supports scheduling
11058 parameters for the EV4, EV5 and EV6 family of processors and will
11059 choose the default values for the instruction set from the processor
11060 you specify. If you do not specify a processor type, GCC will default
11061 to the processor on which the compiler was built.
11063 Supported values for @var{cpu_type} are
11069 Schedules as an EV4 and has no instruction set extensions.
11073 Schedules as an EV5 and has no instruction set extensions.
11077 Schedules as an EV5 and supports the BWX extension.
11082 Schedules as an EV5 and supports the BWX and MAX extensions.
11086 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
11090 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
11093 Native Linux/GNU toolchains also support the value @samp{native},
11094 which selects the best architecture option for the host processor.
11095 @option{-mcpu=native} has no effect if GCC does not recognize
11098 @item -mtune=@var{cpu_type}
11100 Set only the instruction scheduling parameters for machine type
11101 @var{cpu_type}. The instruction set is not changed.
11103 Native Linux/GNU toolchains also support the value @samp{native},
11104 which selects the best architecture option for the host processor.
11105 @option{-mtune=native} has no effect if GCC does not recognize
11108 @item -mmemory-latency=@var{time}
11109 @opindex mmemory-latency
11110 Sets the latency the scheduler should assume for typical memory
11111 references as seen by the application. This number is highly
11112 dependent on the memory access patterns used by the application
11113 and the size of the external cache on the machine.
11115 Valid options for @var{time} are
11119 A decimal number representing clock cycles.
11125 The compiler contains estimates of the number of clock cycles for
11126 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
11127 (also called Dcache, Scache, and Bcache), as well as to main memory.
11128 Note that L3 is only valid for EV5.
11133 @node DEC Alpha/VMS Options
11134 @subsection DEC Alpha/VMS Options
11136 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
11139 @item -mvms-return-codes
11140 @opindex mvms-return-codes
11141 Return VMS condition codes from main. The default is to return POSIX
11142 style condition (e.g.@: error) codes.
11144 @item -mdebug-main=@var{prefix}
11145 @opindex mdebug-main=@var{prefix}
11146 Flag the first routine whose name starts with @var{prefix} as the main
11147 routine for the debugger.
11151 Default to 64bit memory allocation routines.
11155 @subsection FR30 Options
11156 @cindex FR30 Options
11158 These options are defined specifically for the FR30 port.
11162 @item -msmall-model
11163 @opindex msmall-model
11164 Use the small address space model. This can produce smaller code, but
11165 it does assume that all symbolic values and addresses will fit into a
11170 Assume that run-time support has been provided and so there is no need
11171 to include the simulator library (@file{libsim.a}) on the linker
11177 @subsection FRV Options
11178 @cindex FRV Options
11184 Only use the first 32 general purpose registers.
11189 Use all 64 general purpose registers.
11194 Use only the first 32 floating point registers.
11199 Use all 64 floating point registers
11202 @opindex mhard-float
11204 Use hardware instructions for floating point operations.
11207 @opindex msoft-float
11209 Use library routines for floating point operations.
11214 Dynamically allocate condition code registers.
11219 Do not try to dynamically allocate condition code registers, only
11220 use @code{icc0} and @code{fcc0}.
11225 Change ABI to use double word insns.
11230 Do not use double word instructions.
11235 Use floating point double instructions.
11238 @opindex mno-double
11240 Do not use floating point double instructions.
11245 Use media instructions.
11250 Do not use media instructions.
11255 Use multiply and add/subtract instructions.
11258 @opindex mno-muladd
11260 Do not use multiply and add/subtract instructions.
11265 Select the FDPIC ABI, that uses function descriptors to represent
11266 pointers to functions. Without any PIC/PIE-related options, it
11267 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
11268 assumes GOT entries and small data are within a 12-bit range from the
11269 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
11270 are computed with 32 bits.
11271 With a @samp{bfin-elf} target, this option implies @option{-msim}.
11274 @opindex minline-plt
11276 Enable inlining of PLT entries in function calls to functions that are
11277 not known to bind locally. It has no effect without @option{-mfdpic}.
11278 It's enabled by default if optimizing for speed and compiling for
11279 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
11280 optimization option such as @option{-O3} or above is present in the
11286 Assume a large TLS segment when generating thread-local code.
11291 Do not assume a large TLS segment when generating thread-local code.
11296 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
11297 that is known to be in read-only sections. It's enabled by default,
11298 except for @option{-fpic} or @option{-fpie}: even though it may help
11299 make the global offset table smaller, it trades 1 instruction for 4.
11300 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
11301 one of which may be shared by multiple symbols, and it avoids the need
11302 for a GOT entry for the referenced symbol, so it's more likely to be a
11303 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
11305 @item -multilib-library-pic
11306 @opindex multilib-library-pic
11308 Link with the (library, not FD) pic libraries. It's implied by
11309 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
11310 @option{-fpic} without @option{-mfdpic}. You should never have to use
11314 @opindex mlinked-fp
11316 Follow the EABI requirement of always creating a frame pointer whenever
11317 a stack frame is allocated. This option is enabled by default and can
11318 be disabled with @option{-mno-linked-fp}.
11321 @opindex mlong-calls
11323 Use indirect addressing to call functions outside the current
11324 compilation unit. This allows the functions to be placed anywhere
11325 within the 32-bit address space.
11327 @item -malign-labels
11328 @opindex malign-labels
11330 Try to align labels to an 8-byte boundary by inserting nops into the
11331 previous packet. This option only has an effect when VLIW packing
11332 is enabled. It doesn't create new packets; it merely adds nops to
11335 @item -mlibrary-pic
11336 @opindex mlibrary-pic
11338 Generate position-independent EABI code.
11343 Use only the first four media accumulator registers.
11348 Use all eight media accumulator registers.
11353 Pack VLIW instructions.
11358 Do not pack VLIW instructions.
11361 @opindex mno-eflags
11363 Do not mark ABI switches in e_flags.
11366 @opindex mcond-move
11368 Enable the use of conditional-move instructions (default).
11370 This switch is mainly for debugging the compiler and will likely be removed
11371 in a future version.
11373 @item -mno-cond-move
11374 @opindex mno-cond-move
11376 Disable the use of conditional-move instructions.
11378 This switch is mainly for debugging the compiler and will likely be removed
11379 in a future version.
11384 Enable the use of conditional set instructions (default).
11386 This switch is mainly for debugging the compiler and will likely be removed
11387 in a future version.
11392 Disable the use of conditional set instructions.
11394 This switch is mainly for debugging the compiler and will likely be removed
11395 in a future version.
11398 @opindex mcond-exec
11400 Enable the use of conditional execution (default).
11402 This switch is mainly for debugging the compiler and will likely be removed
11403 in a future version.
11405 @item -mno-cond-exec
11406 @opindex mno-cond-exec
11408 Disable the use of conditional execution.
11410 This switch is mainly for debugging the compiler and will likely be removed
11411 in a future version.
11413 @item -mvliw-branch
11414 @opindex mvliw-branch
11416 Run a pass to pack branches into VLIW instructions (default).
11418 This switch is mainly for debugging the compiler and will likely be removed
11419 in a future version.
11421 @item -mno-vliw-branch
11422 @opindex mno-vliw-branch
11424 Do not run a pass to pack branches into VLIW instructions.
11426 This switch is mainly for debugging the compiler and will likely be removed
11427 in a future version.
11429 @item -mmulti-cond-exec
11430 @opindex mmulti-cond-exec
11432 Enable optimization of @code{&&} and @code{||} in conditional execution
11435 This switch is mainly for debugging the compiler and will likely be removed
11436 in a future version.
11438 @item -mno-multi-cond-exec
11439 @opindex mno-multi-cond-exec
11441 Disable optimization of @code{&&} and @code{||} in conditional execution.
11443 This switch is mainly for debugging the compiler and will likely be removed
11444 in a future version.
11446 @item -mnested-cond-exec
11447 @opindex mnested-cond-exec
11449 Enable nested conditional execution optimizations (default).
11451 This switch is mainly for debugging the compiler and will likely be removed
11452 in a future version.
11454 @item -mno-nested-cond-exec
11455 @opindex mno-nested-cond-exec
11457 Disable nested conditional execution optimizations.
11459 This switch is mainly for debugging the compiler and will likely be removed
11460 in a future version.
11462 @item -moptimize-membar
11463 @opindex moptimize-membar
11465 This switch removes redundant @code{membar} instructions from the
11466 compiler generated code. It is enabled by default.
11468 @item -mno-optimize-membar
11469 @opindex mno-optimize-membar
11471 This switch disables the automatic removal of redundant @code{membar}
11472 instructions from the generated code.
11474 @item -mtomcat-stats
11475 @opindex mtomcat-stats
11477 Cause gas to print out tomcat statistics.
11479 @item -mcpu=@var{cpu}
11482 Select the processor type for which to generate code. Possible values are
11483 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
11484 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
11488 @node GNU/Linux Options
11489 @subsection GNU/Linux Options
11491 These @samp{-m} options are defined for GNU/Linux targets:
11496 Use the GNU C library. This is the default except
11497 on @samp{*-*-linux-*uclibc*} and @samp{*-*-linux-*android*} targets.
11501 Use uClibc C library. This is the default on
11502 @samp{*-*-linux-*uclibc*} targets.
11506 Use Bionic C library. This is the default on
11507 @samp{*-*-linux-*android*} targets.
11511 Compile code compatible with Android platform. This is the default on
11512 @samp{*-*-linux-*android*} targets.
11514 When compiling, this option enables @option{-mbionic}, @option{-fPIC},
11515 @option{-fno-exceptions} and @option{-fno-rtti} by default. When linking,
11516 this option makes the GCC driver pass Android-specific options to the linker.
11517 Finally, this option causes the preprocessor macro @code{__ANDROID__}
11520 @item -tno-android-cc
11521 @opindex tno-android-cc
11522 Disable compilation effects of @option{-mandroid}, i.e., do not enable
11523 @option{-mbionic}, @option{-fPIC}, @option{-fno-exceptions} and
11524 @option{-fno-rtti} by default.
11526 @item -tno-android-ld
11527 @opindex tno-android-ld
11528 Disable linking effects of @option{-mandroid}, i.e., pass standard Linux
11529 linking options to the linker.
11533 @node H8/300 Options
11534 @subsection H8/300 Options
11536 These @samp{-m} options are defined for the H8/300 implementations:
11541 Shorten some address references at link time, when possible; uses the
11542 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
11543 ld, Using ld}, for a fuller description.
11547 Generate code for the H8/300H@.
11551 Generate code for the H8S@.
11555 Generate code for the H8S and H8/300H in the normal mode. This switch
11556 must be used either with @option{-mh} or @option{-ms}.
11560 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
11564 Make @code{int} data 32 bits by default.
11567 @opindex malign-300
11568 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
11569 The default for the H8/300H and H8S is to align longs and floats on 4
11571 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
11572 This option has no effect on the H8/300.
11576 @subsection HPPA Options
11577 @cindex HPPA Options
11579 These @samp{-m} options are defined for the HPPA family of computers:
11582 @item -march=@var{architecture-type}
11584 Generate code for the specified architecture. The choices for
11585 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
11586 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
11587 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
11588 architecture option for your machine. Code compiled for lower numbered
11589 architectures will run on higher numbered architectures, but not the
11592 @item -mpa-risc-1-0
11593 @itemx -mpa-risc-1-1
11594 @itemx -mpa-risc-2-0
11595 @opindex mpa-risc-1-0
11596 @opindex mpa-risc-1-1
11597 @opindex mpa-risc-2-0
11598 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
11601 @opindex mbig-switch
11602 Generate code suitable for big switch tables. Use this option only if
11603 the assembler/linker complain about out of range branches within a switch
11606 @item -mjump-in-delay
11607 @opindex mjump-in-delay
11608 Fill delay slots of function calls with unconditional jump instructions
11609 by modifying the return pointer for the function call to be the target
11610 of the conditional jump.
11612 @item -mdisable-fpregs
11613 @opindex mdisable-fpregs
11614 Prevent floating point registers from being used in any manner. This is
11615 necessary for compiling kernels which perform lazy context switching of
11616 floating point registers. If you use this option and attempt to perform
11617 floating point operations, the compiler will abort.
11619 @item -mdisable-indexing
11620 @opindex mdisable-indexing
11621 Prevent the compiler from using indexing address modes. This avoids some
11622 rather obscure problems when compiling MIG generated code under MACH@.
11624 @item -mno-space-regs
11625 @opindex mno-space-regs
11626 Generate code that assumes the target has no space registers. This allows
11627 GCC to generate faster indirect calls and use unscaled index address modes.
11629 Such code is suitable for level 0 PA systems and kernels.
11631 @item -mfast-indirect-calls
11632 @opindex mfast-indirect-calls
11633 Generate code that assumes calls never cross space boundaries. This
11634 allows GCC to emit code which performs faster indirect calls.
11636 This option will not work in the presence of shared libraries or nested
11639 @item -mfixed-range=@var{register-range}
11640 @opindex mfixed-range
11641 Generate code treating the given register range as fixed registers.
11642 A fixed register is one that the register allocator can not use. This is
11643 useful when compiling kernel code. A register range is specified as
11644 two registers separated by a dash. Multiple register ranges can be
11645 specified separated by a comma.
11647 @item -mlong-load-store
11648 @opindex mlong-load-store
11649 Generate 3-instruction load and store sequences as sometimes required by
11650 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
11653 @item -mportable-runtime
11654 @opindex mportable-runtime
11655 Use the portable calling conventions proposed by HP for ELF systems.
11659 Enable the use of assembler directives only GAS understands.
11661 @item -mschedule=@var{cpu-type}
11663 Schedule code according to the constraints for the machine type
11664 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
11665 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
11666 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
11667 proper scheduling option for your machine. The default scheduling is
11671 @opindex mlinker-opt
11672 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
11673 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
11674 linkers in which they give bogus error messages when linking some programs.
11677 @opindex msoft-float
11678 Generate output containing library calls for floating point.
11679 @strong{Warning:} the requisite libraries are not available for all HPPA
11680 targets. Normally the facilities of the machine's usual C compiler are
11681 used, but this cannot be done directly in cross-compilation. You must make
11682 your own arrangements to provide suitable library functions for
11685 @option{-msoft-float} changes the calling convention in the output file;
11686 therefore, it is only useful if you compile @emph{all} of a program with
11687 this option. In particular, you need to compile @file{libgcc.a}, the
11688 library that comes with GCC, with @option{-msoft-float} in order for
11693 Generate the predefine, @code{_SIO}, for server IO@. The default is
11694 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
11695 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
11696 options are available under HP-UX and HI-UX@.
11700 Use GNU ld specific options. This passes @option{-shared} to ld when
11701 building a shared library. It is the default when GCC is configured,
11702 explicitly or implicitly, with the GNU linker. This option does not
11703 have any affect on which ld is called, it only changes what parameters
11704 are passed to that ld. The ld that is called is determined by the
11705 @option{--with-ld} configure option, GCC's program search path, and
11706 finally by the user's @env{PATH}. The linker used by GCC can be printed
11707 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
11708 on the 64 bit HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
11712 Use HP ld specific options. This passes @option{-b} to ld when building
11713 a shared library and passes @option{+Accept TypeMismatch} to ld on all
11714 links. It is the default when GCC is configured, explicitly or
11715 implicitly, with the HP linker. This option does not have any affect on
11716 which ld is called, it only changes what parameters are passed to that
11717 ld. The ld that is called is determined by the @option{--with-ld}
11718 configure option, GCC's program search path, and finally by the user's
11719 @env{PATH}. The linker used by GCC can be printed using @samp{which
11720 `gcc -print-prog-name=ld`}. This option is only available on the 64 bit
11721 HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
11724 @opindex mno-long-calls
11725 Generate code that uses long call sequences. This ensures that a call
11726 is always able to reach linker generated stubs. The default is to generate
11727 long calls only when the distance from the call site to the beginning
11728 of the function or translation unit, as the case may be, exceeds a
11729 predefined limit set by the branch type being used. The limits for
11730 normal calls are 7,600,000 and 240,000 bytes, respectively for the
11731 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
11734 Distances are measured from the beginning of functions when using the
11735 @option{-ffunction-sections} option, or when using the @option{-mgas}
11736 and @option{-mno-portable-runtime} options together under HP-UX with
11739 It is normally not desirable to use this option as it will degrade
11740 performance. However, it may be useful in large applications,
11741 particularly when partial linking is used to build the application.
11743 The types of long calls used depends on the capabilities of the
11744 assembler and linker, and the type of code being generated. The
11745 impact on systems that support long absolute calls, and long pic
11746 symbol-difference or pc-relative calls should be relatively small.
11747 However, an indirect call is used on 32-bit ELF systems in pic code
11748 and it is quite long.
11750 @item -munix=@var{unix-std}
11752 Generate compiler predefines and select a startfile for the specified
11753 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
11754 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
11755 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
11756 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
11757 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
11760 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
11761 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
11762 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
11763 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
11764 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
11765 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
11767 It is @emph{important} to note that this option changes the interfaces
11768 for various library routines. It also affects the operational behavior
11769 of the C library. Thus, @emph{extreme} care is needed in using this
11772 Library code that is intended to operate with more than one UNIX
11773 standard must test, set and restore the variable @var{__xpg4_extended_mask}
11774 as appropriate. Most GNU software doesn't provide this capability.
11778 Suppress the generation of link options to search libdld.sl when the
11779 @option{-static} option is specified on HP-UX 10 and later.
11783 The HP-UX implementation of setlocale in libc has a dependency on
11784 libdld.sl. There isn't an archive version of libdld.sl. Thus,
11785 when the @option{-static} option is specified, special link options
11786 are needed to resolve this dependency.
11788 On HP-UX 10 and later, the GCC driver adds the necessary options to
11789 link with libdld.sl when the @option{-static} option is specified.
11790 This causes the resulting binary to be dynamic. On the 64-bit port,
11791 the linkers generate dynamic binaries by default in any case. The
11792 @option{-nolibdld} option can be used to prevent the GCC driver from
11793 adding these link options.
11797 Add support for multithreading with the @dfn{dce thread} library
11798 under HP-UX@. This option sets flags for both the preprocessor and
11802 @node i386 and x86-64 Options
11803 @subsection Intel 386 and AMD x86-64 Options
11804 @cindex i386 Options
11805 @cindex x86-64 Options
11806 @cindex Intel 386 Options
11807 @cindex AMD x86-64 Options
11809 These @samp{-m} options are defined for the i386 and x86-64 family of
11813 @item -mtune=@var{cpu-type}
11815 Tune to @var{cpu-type} everything applicable about the generated code, except
11816 for the ABI and the set of available instructions. The choices for
11817 @var{cpu-type} are:
11820 Produce code optimized for the most common IA32/AMD64/EM64T processors.
11821 If you know the CPU on which your code will run, then you should use
11822 the corresponding @option{-mtune} option instead of
11823 @option{-mtune=generic}. But, if you do not know exactly what CPU users
11824 of your application will have, then you should use this option.
11826 As new processors are deployed in the marketplace, the behavior of this
11827 option will change. Therefore, if you upgrade to a newer version of
11828 GCC, the code generated option will change to reflect the processors
11829 that were most common when that version of GCC was released.
11831 There is no @option{-march=generic} option because @option{-march}
11832 indicates the instruction set the compiler can use, and there is no
11833 generic instruction set applicable to all processors. In contrast,
11834 @option{-mtune} indicates the processor (or, in this case, collection of
11835 processors) for which the code is optimized.
11837 This selects the CPU to tune for at compilation time by determining
11838 the processor type of the compiling machine. Using @option{-mtune=native}
11839 will produce code optimized for the local machine under the constraints
11840 of the selected instruction set. Using @option{-march=native} will
11841 enable all instruction subsets supported by the local machine (hence
11842 the result might not run on different machines).
11844 Original Intel's i386 CPU@.
11846 Intel's i486 CPU@. (No scheduling is implemented for this chip.)
11847 @item i586, pentium
11848 Intel Pentium CPU with no MMX support.
11850 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
11852 Intel PentiumPro CPU@.
11854 Same as @code{generic}, but when used as @code{march} option, PentiumPro
11855 instruction set will be used, so the code will run on all i686 family chips.
11857 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
11858 @item pentium3, pentium3m
11859 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
11862 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
11863 support. Used by Centrino notebooks.
11864 @item pentium4, pentium4m
11865 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
11867 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
11870 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
11871 SSE2 and SSE3 instruction set support.
11873 Intel Core2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
11874 instruction set support.
11876 Intel Atom CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
11877 instruction set support.
11879 AMD K6 CPU with MMX instruction set support.
11881 Improved versions of AMD K6 CPU with MMX and 3DNow!@: instruction set support.
11882 @item athlon, athlon-tbird
11883 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3DNow!@: and SSE prefetch instructions
11885 @item athlon-4, athlon-xp, athlon-mp
11886 Improved AMD Athlon CPU with MMX, 3DNow!, enhanced 3DNow!@: and full SSE
11887 instruction set support.
11888 @item k8, opteron, athlon64, athlon-fx
11889 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
11890 MMX, SSE, SSE2, 3DNow!, enhanced 3DNow!@: and 64-bit instruction set extensions.)
11891 @item k8-sse3, opteron-sse3, athlon64-sse3
11892 Improved versions of k8, opteron and athlon64 with SSE3 instruction set support.
11893 @item amdfam10, barcelona
11894 AMD Family 10h core based CPUs with x86-64 instruction set support. (This
11895 supersets MMX, SSE, SSE2, SSE3, SSE4A, 3DNow!, enhanced 3DNow!, ABM and 64-bit
11896 instruction set extensions.)
11898 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
11901 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3DNow!@:
11902 instruction set support.
11904 Via C3 CPU with MMX and 3DNow!@: instruction set support. (No scheduling is
11905 implemented for this chip.)
11907 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
11908 implemented for this chip.)
11910 Embedded AMD CPU with MMX and 3DNow!@: instruction set support.
11913 While picking a specific @var{cpu-type} will schedule things appropriately
11914 for that particular chip, the compiler will not generate any code that
11915 does not run on the i386 without the @option{-march=@var{cpu-type}} option
11918 @item -march=@var{cpu-type}
11920 Generate instructions for the machine type @var{cpu-type}. The choices
11921 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
11922 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
11924 @item -mcpu=@var{cpu-type}
11926 A deprecated synonym for @option{-mtune}.
11928 @item -mfpmath=@var{unit}
11930 Generate floating point arithmetics for selected unit @var{unit}. The choices
11931 for @var{unit} are:
11935 Use the standard 387 floating point coprocessor present majority of chips and
11936 emulated otherwise. Code compiled with this option will run almost everywhere.
11937 The temporary results are computed in 80bit precision instead of precision
11938 specified by the type resulting in slightly different results compared to most
11939 of other chips. See @option{-ffloat-store} for more detailed description.
11941 This is the default choice for i386 compiler.
11944 Use scalar floating point instructions present in the SSE instruction set.
11945 This instruction set is supported by Pentium3 and newer chips, in the AMD line
11946 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
11947 instruction set supports only single precision arithmetics, thus the double and
11948 extended precision arithmetics is still done using 387. Later version, present
11949 only in Pentium4 and the future AMD x86-64 chips supports double precision
11952 For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse}
11953 or @option{-msse2} switches to enable SSE extensions and make this option
11954 effective. For the x86-64 compiler, these extensions are enabled by default.
11956 The resulting code should be considerably faster in the majority of cases and avoid
11957 the numerical instability problems of 387 code, but may break some existing
11958 code that expects temporaries to be 80bit.
11960 This is the default choice for the x86-64 compiler.
11965 Attempt to utilize both instruction sets at once. This effectively double the
11966 amount of available registers and on chips with separate execution units for
11967 387 and SSE the execution resources too. Use this option with care, as it is
11968 still experimental, because the GCC register allocator does not model separate
11969 functional units well resulting in instable performance.
11972 @item -masm=@var{dialect}
11973 @opindex masm=@var{dialect}
11974 Output asm instructions using selected @var{dialect}. Supported
11975 choices are @samp{intel} or @samp{att} (the default one). Darwin does
11976 not support @samp{intel}.
11979 @itemx -mno-ieee-fp
11981 @opindex mno-ieee-fp
11982 Control whether or not the compiler uses IEEE floating point
11983 comparisons. These handle correctly the case where the result of a
11984 comparison is unordered.
11987 @opindex msoft-float
11988 Generate output containing library calls for floating point.
11989 @strong{Warning:} the requisite libraries are not part of GCC@.
11990 Normally the facilities of the machine's usual C compiler are used, but
11991 this can't be done directly in cross-compilation. You must make your
11992 own arrangements to provide suitable library functions for
11995 On machines where a function returns floating point results in the 80387
11996 register stack, some floating point opcodes may be emitted even if
11997 @option{-msoft-float} is used.
11999 @item -mno-fp-ret-in-387
12000 @opindex mno-fp-ret-in-387
12001 Do not use the FPU registers for return values of functions.
12003 The usual calling convention has functions return values of types
12004 @code{float} and @code{double} in an FPU register, even if there
12005 is no FPU@. The idea is that the operating system should emulate
12008 The option @option{-mno-fp-ret-in-387} causes such values to be returned
12009 in ordinary CPU registers instead.
12011 @item -mno-fancy-math-387
12012 @opindex mno-fancy-math-387
12013 Some 387 emulators do not support the @code{sin}, @code{cos} and
12014 @code{sqrt} instructions for the 387. Specify this option to avoid
12015 generating those instructions. This option is the default on FreeBSD,
12016 OpenBSD and NetBSD@. This option is overridden when @option{-march}
12017 indicates that the target cpu will always have an FPU and so the
12018 instruction will not need emulation. As of revision 2.6.1, these
12019 instructions are not generated unless you also use the
12020 @option{-funsafe-math-optimizations} switch.
12022 @item -malign-double
12023 @itemx -mno-align-double
12024 @opindex malign-double
12025 @opindex mno-align-double
12026 Control whether GCC aligns @code{double}, @code{long double}, and
12027 @code{long long} variables on a two word boundary or a one word
12028 boundary. Aligning @code{double} variables on a two word boundary will
12029 produce code that runs somewhat faster on a @samp{Pentium} at the
12030 expense of more memory.
12032 On x86-64, @option{-malign-double} is enabled by default.
12034 @strong{Warning:} if you use the @option{-malign-double} switch,
12035 structures containing the above types will be aligned differently than
12036 the published application binary interface specifications for the 386
12037 and will not be binary compatible with structures in code compiled
12038 without that switch.
12040 @item -m96bit-long-double
12041 @itemx -m128bit-long-double
12042 @opindex m96bit-long-double
12043 @opindex m128bit-long-double
12044 These switches control the size of @code{long double} type. The i386
12045 application binary interface specifies the size to be 96 bits,
12046 so @option{-m96bit-long-double} is the default in 32 bit mode.
12048 Modern architectures (Pentium and newer) would prefer @code{long double}
12049 to be aligned to an 8 or 16 byte boundary. In arrays or structures
12050 conforming to the ABI, this would not be possible. So specifying a
12051 @option{-m128bit-long-double} will align @code{long double}
12052 to a 16 byte boundary by padding the @code{long double} with an additional
12055 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
12056 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
12058 Notice that neither of these options enable any extra precision over the x87
12059 standard of 80 bits for a @code{long double}.
12061 @strong{Warning:} if you override the default value for your target ABI, the
12062 structures and arrays containing @code{long double} variables will change
12063 their size as well as function calling convention for function taking
12064 @code{long double} will be modified. Hence they will not be binary
12065 compatible with arrays or structures in code compiled without that switch.
12067 @item -mlarge-data-threshold=@var{number}
12068 @opindex mlarge-data-threshold=@var{number}
12069 When @option{-mcmodel=medium} is specified, the data greater than
12070 @var{threshold} are placed in large data section. This value must be the
12071 same across all object linked into the binary and defaults to 65535.
12075 Use a different function-calling convention, in which functions that
12076 take a fixed number of arguments return with the @code{ret} @var{num}
12077 instruction, which pops their arguments while returning. This saves one
12078 instruction in the caller since there is no need to pop the arguments
12081 You can specify that an individual function is called with this calling
12082 sequence with the function attribute @samp{stdcall}. You can also
12083 override the @option{-mrtd} option by using the function attribute
12084 @samp{cdecl}. @xref{Function Attributes}.
12086 @strong{Warning:} this calling convention is incompatible with the one
12087 normally used on Unix, so you cannot use it if you need to call
12088 libraries compiled with the Unix compiler.
12090 Also, you must provide function prototypes for all functions that
12091 take variable numbers of arguments (including @code{printf});
12092 otherwise incorrect code will be generated for calls to those
12095 In addition, seriously incorrect code will result if you call a
12096 function with too many arguments. (Normally, extra arguments are
12097 harmlessly ignored.)
12099 @item -mregparm=@var{num}
12101 Control how many registers are used to pass integer arguments. By
12102 default, no registers are used to pass arguments, and at most 3
12103 registers can be used. You can control this behavior for a specific
12104 function by using the function attribute @samp{regparm}.
12105 @xref{Function Attributes}.
12107 @strong{Warning:} if you use this switch, and
12108 @var{num} is nonzero, then you must build all modules with the same
12109 value, including any libraries. This includes the system libraries and
12113 @opindex msseregparm
12114 Use SSE register passing conventions for float and double arguments
12115 and return values. You can control this behavior for a specific
12116 function by using the function attribute @samp{sseregparm}.
12117 @xref{Function Attributes}.
12119 @strong{Warning:} if you use this switch then you must build all
12120 modules with the same value, including any libraries. This includes
12121 the system libraries and startup modules.
12130 Set 80387 floating-point precision to 32, 64 or 80 bits. When @option{-mpc32}
12131 is specified, the significands of results of floating-point operations are
12132 rounded to 24 bits (single precision); @option{-mpc64} rounds the
12133 significands of results of floating-point operations to 53 bits (double
12134 precision) and @option{-mpc80} rounds the significands of results of
12135 floating-point operations to 64 bits (extended double precision), which is
12136 the default. When this option is used, floating-point operations in higher
12137 precisions are not available to the programmer without setting the FPU
12138 control word explicitly.
12140 Setting the rounding of floating-point operations to less than the default
12141 80 bits can speed some programs by 2% or more. Note that some mathematical
12142 libraries assume that extended precision (80 bit) floating-point operations
12143 are enabled by default; routines in such libraries could suffer significant
12144 loss of accuracy, typically through so-called "catastrophic cancellation",
12145 when this option is used to set the precision to less than extended precision.
12147 @item -mstackrealign
12148 @opindex mstackrealign
12149 Realign the stack at entry. On the Intel x86, the @option{-mstackrealign}
12150 option will generate an alternate prologue and epilogue that realigns the
12151 runtime stack if necessary. This supports mixing legacy codes that keep
12152 a 4-byte aligned stack with modern codes that keep a 16-byte stack for
12153 SSE compatibility. See also the attribute @code{force_align_arg_pointer},
12154 applicable to individual functions.
12156 @item -mpreferred-stack-boundary=@var{num}
12157 @opindex mpreferred-stack-boundary
12158 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
12159 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
12160 the default is 4 (16 bytes or 128 bits).
12162 @item -mincoming-stack-boundary=@var{num}
12163 @opindex mincoming-stack-boundary
12164 Assume the incoming stack is aligned to a 2 raised to @var{num} byte
12165 boundary. If @option{-mincoming-stack-boundary} is not specified,
12166 the one specified by @option{-mpreferred-stack-boundary} will be used.
12168 On Pentium and PentiumPro, @code{double} and @code{long double} values
12169 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
12170 suffer significant run time performance penalties. On Pentium III, the
12171 Streaming SIMD Extension (SSE) data type @code{__m128} may not work
12172 properly if it is not 16 byte aligned.
12174 To ensure proper alignment of this values on the stack, the stack boundary
12175 must be as aligned as that required by any value stored on the stack.
12176 Further, every function must be generated such that it keeps the stack
12177 aligned. Thus calling a function compiled with a higher preferred
12178 stack boundary from a function compiled with a lower preferred stack
12179 boundary will most likely misalign the stack. It is recommended that
12180 libraries that use callbacks always use the default setting.
12182 This extra alignment does consume extra stack space, and generally
12183 increases code size. Code that is sensitive to stack space usage, such
12184 as embedded systems and operating system kernels, may want to reduce the
12185 preferred alignment to @option{-mpreferred-stack-boundary=2}.
12229 These switches enable or disable the use of instructions in the MMX,
12230 SSE, SSE2, SSE3, SSSE3, SSE4.1, AVX, AES, PCLMUL, SSE4A, FMA4, XOP,
12231 LWP, ABM or 3DNow!@: extended instruction sets.
12232 These extensions are also available as built-in functions: see
12233 @ref{X86 Built-in Functions}, for details of the functions enabled and
12234 disabled by these switches.
12236 To have SSE/SSE2 instructions generated automatically from floating-point
12237 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
12239 GCC depresses SSEx instructions when @option{-mavx} is used. Instead, it
12240 generates new AVX instructions or AVX equivalence for all SSEx instructions
12243 These options will enable GCC to use these extended instructions in
12244 generated code, even without @option{-mfpmath=sse}. Applications which
12245 perform runtime CPU detection must compile separate files for each
12246 supported architecture, using the appropriate flags. In particular,
12247 the file containing the CPU detection code should be compiled without
12251 @itemx -mno-fused-madd
12252 @opindex mfused-madd
12253 @opindex mno-fused-madd
12254 Do (don't) generate code that uses the fused multiply/add or multiply/subtract
12255 instructions. The default is to use these instructions.
12259 This option instructs GCC to emit a @code{cld} instruction in the prologue
12260 of functions that use string instructions. String instructions depend on
12261 the DF flag to select between autoincrement or autodecrement mode. While the
12262 ABI specifies the DF flag to be cleared on function entry, some operating
12263 systems violate this specification by not clearing the DF flag in their
12264 exception dispatchers. The exception handler can be invoked with the DF flag
12265 set which leads to wrong direction mode, when string instructions are used.
12266 This option can be enabled by default on 32-bit x86 targets by configuring
12267 GCC with the @option{--enable-cld} configure option. Generation of @code{cld}
12268 instructions can be suppressed with the @option{-mno-cld} compiler option
12273 This option will enable GCC to use CMPXCHG16B instruction in generated code.
12274 CMPXCHG16B allows for atomic operations on 128-bit double quadword (or oword)
12275 data types. This is useful for high resolution counters that could be updated
12276 by multiple processors (or cores). This instruction is generated as part of
12277 atomic built-in functions: see @ref{Atomic Builtins} for details.
12281 This option will enable GCC to use SAHF instruction in generated 64-bit code.
12282 Early Intel CPUs with Intel 64 lacked LAHF and SAHF instructions supported
12283 by AMD64 until introduction of Pentium 4 G1 step in December 2005. LAHF and
12284 SAHF are load and store instructions, respectively, for certain status flags.
12285 In 64-bit mode, SAHF instruction is used to optimize @code{fmod}, @code{drem}
12286 or @code{remainder} built-in functions: see @ref{Other Builtins} for details.
12290 This option will enable GCC to use movbe instruction to implement
12291 @code{__builtin_bswap32} and @code{__builtin_bswap64}.
12295 This option will enable built-in functions, @code{__builtin_ia32_crc32qi},
12296 @code{__builtin_ia32_crc32hi}. @code{__builtin_ia32_crc32si} and
12297 @code{__builtin_ia32_crc32di} to generate the crc32 machine instruction.
12301 This option will enable GCC to use RCPSS and RSQRTSS instructions (and their
12302 vectorized variants RCPPS and RSQRTPS) with an additional Newton-Raphson step
12303 to increase precision instead of DIVSS and SQRTSS (and their vectorized
12304 variants) for single precision floating point arguments. These instructions
12305 are generated only when @option{-funsafe-math-optimizations} is enabled
12306 together with @option{-finite-math-only} and @option{-fno-trapping-math}.
12307 Note that while the throughput of the sequence is higher than the throughput
12308 of the non-reciprocal instruction, the precision of the sequence can be
12309 decreased by up to 2 ulp (i.e. the inverse of 1.0 equals 0.99999994).
12311 Note that GCC implements 1.0f/sqrtf(x) in terms of RSQRTSS (or RSQRTPS)
12312 already with @option{-ffast-math} (or the above option combination), and
12313 doesn't need @option{-mrecip}.
12315 @item -mveclibabi=@var{type}
12316 @opindex mveclibabi
12317 Specifies the ABI type to use for vectorizing intrinsics using an
12318 external library. Supported types are @code{svml} for the Intel short
12319 vector math library and @code{acml} for the AMD math core library style
12320 of interfacing. GCC will currently emit calls to @code{vmldExp2},
12321 @code{vmldLn2}, @code{vmldLog102}, @code{vmldLog102}, @code{vmldPow2},
12322 @code{vmldTanh2}, @code{vmldTan2}, @code{vmldAtan2}, @code{vmldAtanh2},
12323 @code{vmldCbrt2}, @code{vmldSinh2}, @code{vmldSin2}, @code{vmldAsinh2},
12324 @code{vmldAsin2}, @code{vmldCosh2}, @code{vmldCos2}, @code{vmldAcosh2},
12325 @code{vmldAcos2}, @code{vmlsExp4}, @code{vmlsLn4}, @code{vmlsLog104},
12326 @code{vmlsLog104}, @code{vmlsPow4}, @code{vmlsTanh4}, @code{vmlsTan4},
12327 @code{vmlsAtan4}, @code{vmlsAtanh4}, @code{vmlsCbrt4}, @code{vmlsSinh4},
12328 @code{vmlsSin4}, @code{vmlsAsinh4}, @code{vmlsAsin4}, @code{vmlsCosh4},
12329 @code{vmlsCos4}, @code{vmlsAcosh4} and @code{vmlsAcos4} for corresponding
12330 function type when @option{-mveclibabi=svml} is used and @code{__vrd2_sin},
12331 @code{__vrd2_cos}, @code{__vrd2_exp}, @code{__vrd2_log}, @code{__vrd2_log2},
12332 @code{__vrd2_log10}, @code{__vrs4_sinf}, @code{__vrs4_cosf},
12333 @code{__vrs4_expf}, @code{__vrs4_logf}, @code{__vrs4_log2f},
12334 @code{__vrs4_log10f} and @code{__vrs4_powf} for corresponding function type
12335 when @option{-mveclibabi=acml} is used. Both @option{-ftree-vectorize} and
12336 @option{-funsafe-math-optimizations} have to be enabled. A SVML or ACML ABI
12337 compatible library will have to be specified at link time.
12339 @item -mabi=@var{name}
12341 Generate code for the specified calling convention. Permissible values
12342 are: @samp{sysv} for the ABI used on GNU/Linux and other systems and
12343 @samp{ms} for the Microsoft ABI. The default is to use the Microsoft
12344 ABI when targeting Windows. On all other systems, the default is the
12345 SYSV ABI. You can control this behavior for a specific function by
12346 using the function attribute @samp{ms_abi}/@samp{sysv_abi}.
12347 @xref{Function Attributes}.
12350 @itemx -mno-push-args
12351 @opindex mpush-args
12352 @opindex mno-push-args
12353 Use PUSH operations to store outgoing parameters. This method is shorter
12354 and usually equally fast as method using SUB/MOV operations and is enabled
12355 by default. In some cases disabling it may improve performance because of
12356 improved scheduling and reduced dependencies.
12358 @item -maccumulate-outgoing-args
12359 @opindex maccumulate-outgoing-args
12360 If enabled, the maximum amount of space required for outgoing arguments will be
12361 computed in the function prologue. This is faster on most modern CPUs
12362 because of reduced dependencies, improved scheduling and reduced stack usage
12363 when preferred stack boundary is not equal to 2. The drawback is a notable
12364 increase in code size. This switch implies @option{-mno-push-args}.
12368 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
12369 on thread-safe exception handling must compile and link all code with the
12370 @option{-mthreads} option. When compiling, @option{-mthreads} defines
12371 @option{-D_MT}; when linking, it links in a special thread helper library
12372 @option{-lmingwthrd} which cleans up per thread exception handling data.
12374 @item -mno-align-stringops
12375 @opindex mno-align-stringops
12376 Do not align destination of inlined string operations. This switch reduces
12377 code size and improves performance in case the destination is already aligned,
12378 but GCC doesn't know about it.
12380 @item -minline-all-stringops
12381 @opindex minline-all-stringops
12382 By default GCC inlines string operations only when destination is known to be
12383 aligned at least to 4 byte boundary. This enables more inlining, increase code
12384 size, but may improve performance of code that depends on fast memcpy, strlen
12385 and memset for short lengths.
12387 @item -minline-stringops-dynamically
12388 @opindex minline-stringops-dynamically
12389 For string operation of unknown size, inline runtime checks so for small
12390 blocks inline code is used, while for large blocks library call is used.
12392 @item -mstringop-strategy=@var{alg}
12393 @opindex mstringop-strategy=@var{alg}
12394 Overwrite internal decision heuristic about particular algorithm to inline
12395 string operation with. The allowed values are @code{rep_byte},
12396 @code{rep_4byte}, @code{rep_8byte} for expanding using i386 @code{rep} prefix
12397 of specified size, @code{byte_loop}, @code{loop}, @code{unrolled_loop} for
12398 expanding inline loop, @code{libcall} for always expanding library call.
12400 @item -momit-leaf-frame-pointer
12401 @opindex momit-leaf-frame-pointer
12402 Don't keep the frame pointer in a register for leaf functions. This
12403 avoids the instructions to save, set up and restore frame pointers and
12404 makes an extra register available in leaf functions. The option
12405 @option{-fomit-frame-pointer} removes the frame pointer for all functions
12406 which might make debugging harder.
12408 @item -mtls-direct-seg-refs
12409 @itemx -mno-tls-direct-seg-refs
12410 @opindex mtls-direct-seg-refs
12411 Controls whether TLS variables may be accessed with offsets from the
12412 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
12413 or whether the thread base pointer must be added. Whether or not this
12414 is legal depends on the operating system, and whether it maps the
12415 segment to cover the entire TLS area.
12417 For systems that use GNU libc, the default is on.
12420 @itemx -mno-sse2avx
12422 Specify that the assembler should encode SSE instructions with VEX
12423 prefix. The option @option{-mavx} turns this on by default.
12426 These @samp{-m} switches are supported in addition to the above
12427 on AMD x86-64 processors in 64-bit environments.
12434 Generate code for a 32-bit or 64-bit environment.
12435 The 32-bit environment sets int, long and pointer to 32 bits and
12436 generates code that runs on any i386 system.
12437 The 64-bit environment sets int to 32 bits and long and pointer
12438 to 64 bits and generates code for AMD's x86-64 architecture. For
12439 darwin only the -m64 option turns off the @option{-fno-pic} and
12440 @option{-mdynamic-no-pic} options.
12442 @item -mno-red-zone
12443 @opindex mno-red-zone
12444 Do not use a so called red zone for x86-64 code. The red zone is mandated
12445 by the x86-64 ABI, it is a 128-byte area beyond the location of the
12446 stack pointer that will not be modified by signal or interrupt handlers
12447 and therefore can be used for temporary data without adjusting the stack
12448 pointer. The flag @option{-mno-red-zone} disables this red zone.
12450 @item -mcmodel=small
12451 @opindex mcmodel=small
12452 Generate code for the small code model: the program and its symbols must
12453 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
12454 Programs can be statically or dynamically linked. This is the default
12457 @item -mcmodel=kernel
12458 @opindex mcmodel=kernel
12459 Generate code for the kernel code model. The kernel runs in the
12460 negative 2 GB of the address space.
12461 This model has to be used for Linux kernel code.
12463 @item -mcmodel=medium
12464 @opindex mcmodel=medium
12465 Generate code for the medium model: The program is linked in the lower 2
12466 GB of the address space. Small symbols are also placed there. Symbols
12467 with sizes larger than @option{-mlarge-data-threshold} are put into
12468 large data or bss sections and can be located above 2GB. Programs can
12469 be statically or dynamically linked.
12471 @item -mcmodel=large
12472 @opindex mcmodel=large
12473 Generate code for the large model: This model makes no assumptions
12474 about addresses and sizes of sections.
12477 @node IA-64 Options
12478 @subsection IA-64 Options
12479 @cindex IA-64 Options
12481 These are the @samp{-m} options defined for the Intel IA-64 architecture.
12485 @opindex mbig-endian
12486 Generate code for a big endian target. This is the default for HP-UX@.
12488 @item -mlittle-endian
12489 @opindex mlittle-endian
12490 Generate code for a little endian target. This is the default for AIX5
12496 @opindex mno-gnu-as
12497 Generate (or don't) code for the GNU assembler. This is the default.
12498 @c Also, this is the default if the configure option @option{--with-gnu-as}
12504 @opindex mno-gnu-ld
12505 Generate (or don't) code for the GNU linker. This is the default.
12506 @c Also, this is the default if the configure option @option{--with-gnu-ld}
12511 Generate code that does not use a global pointer register. The result
12512 is not position independent code, and violates the IA-64 ABI@.
12514 @item -mvolatile-asm-stop
12515 @itemx -mno-volatile-asm-stop
12516 @opindex mvolatile-asm-stop
12517 @opindex mno-volatile-asm-stop
12518 Generate (or don't) a stop bit immediately before and after volatile asm
12521 @item -mregister-names
12522 @itemx -mno-register-names
12523 @opindex mregister-names
12524 @opindex mno-register-names
12525 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
12526 the stacked registers. This may make assembler output more readable.
12532 Disable (or enable) optimizations that use the small data section. This may
12533 be useful for working around optimizer bugs.
12535 @item -mconstant-gp
12536 @opindex mconstant-gp
12537 Generate code that uses a single constant global pointer value. This is
12538 useful when compiling kernel code.
12542 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
12543 This is useful when compiling firmware code.
12545 @item -minline-float-divide-min-latency
12546 @opindex minline-float-divide-min-latency
12547 Generate code for inline divides of floating point values
12548 using the minimum latency algorithm.
12550 @item -minline-float-divide-max-throughput
12551 @opindex minline-float-divide-max-throughput
12552 Generate code for inline divides of floating point values
12553 using the maximum throughput algorithm.
12555 @item -mno-inline-float-divide
12556 @opindex mno-inline-float-divide
12557 Do not generate inline code for divides of floating point values.
12559 @item -minline-int-divide-min-latency
12560 @opindex minline-int-divide-min-latency
12561 Generate code for inline divides of integer values
12562 using the minimum latency algorithm.
12564 @item -minline-int-divide-max-throughput
12565 @opindex minline-int-divide-max-throughput
12566 Generate code for inline divides of integer values
12567 using the maximum throughput algorithm.
12569 @item -mno-inline-int-divide
12570 @opindex mno-inline-int-divide
12571 Do not generate inline code for divides of integer values.
12573 @item -minline-sqrt-min-latency
12574 @opindex minline-sqrt-min-latency
12575 Generate code for inline square roots
12576 using the minimum latency algorithm.
12578 @item -minline-sqrt-max-throughput
12579 @opindex minline-sqrt-max-throughput
12580 Generate code for inline square roots
12581 using the maximum throughput algorithm.
12583 @item -mno-inline-sqrt
12584 @opindex mno-inline-sqrt
12585 Do not generate inline code for sqrt.
12588 @itemx -mno-fused-madd
12589 @opindex mfused-madd
12590 @opindex mno-fused-madd
12591 Do (don't) generate code that uses the fused multiply/add or multiply/subtract
12592 instructions. The default is to use these instructions.
12594 @item -mno-dwarf2-asm
12595 @itemx -mdwarf2-asm
12596 @opindex mno-dwarf2-asm
12597 @opindex mdwarf2-asm
12598 Don't (or do) generate assembler code for the DWARF2 line number debugging
12599 info. This may be useful when not using the GNU assembler.
12601 @item -mearly-stop-bits
12602 @itemx -mno-early-stop-bits
12603 @opindex mearly-stop-bits
12604 @opindex mno-early-stop-bits
12605 Allow stop bits to be placed earlier than immediately preceding the
12606 instruction that triggered the stop bit. This can improve instruction
12607 scheduling, but does not always do so.
12609 @item -mfixed-range=@var{register-range}
12610 @opindex mfixed-range
12611 Generate code treating the given register range as fixed registers.
12612 A fixed register is one that the register allocator can not use. This is
12613 useful when compiling kernel code. A register range is specified as
12614 two registers separated by a dash. Multiple register ranges can be
12615 specified separated by a comma.
12617 @item -mtls-size=@var{tls-size}
12619 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
12622 @item -mtune=@var{cpu-type}
12624 Tune the instruction scheduling for a particular CPU, Valid values are
12625 itanium, itanium1, merced, itanium2, and mckinley.
12631 Generate code for a 32-bit or 64-bit environment.
12632 The 32-bit environment sets int, long and pointer to 32 bits.
12633 The 64-bit environment sets int to 32 bits and long and pointer
12634 to 64 bits. These are HP-UX specific flags.
12636 @item -mno-sched-br-data-spec
12637 @itemx -msched-br-data-spec
12638 @opindex mno-sched-br-data-spec
12639 @opindex msched-br-data-spec
12640 (Dis/En)able data speculative scheduling before reload.
12641 This will result in generation of the ld.a instructions and
12642 the corresponding check instructions (ld.c / chk.a).
12643 The default is 'disable'.
12645 @item -msched-ar-data-spec
12646 @itemx -mno-sched-ar-data-spec
12647 @opindex msched-ar-data-spec
12648 @opindex mno-sched-ar-data-spec
12649 (En/Dis)able data speculative scheduling after reload.
12650 This will result in generation of the ld.a instructions and
12651 the corresponding check instructions (ld.c / chk.a).
12652 The default is 'enable'.
12654 @item -mno-sched-control-spec
12655 @itemx -msched-control-spec
12656 @opindex mno-sched-control-spec
12657 @opindex msched-control-spec
12658 (Dis/En)able control speculative scheduling. This feature is
12659 available only during region scheduling (i.e.@: before reload).
12660 This will result in generation of the ld.s instructions and
12661 the corresponding check instructions chk.s .
12662 The default is 'disable'.
12664 @item -msched-br-in-data-spec
12665 @itemx -mno-sched-br-in-data-spec
12666 @opindex msched-br-in-data-spec
12667 @opindex mno-sched-br-in-data-spec
12668 (En/Dis)able speculative scheduling of the instructions that
12669 are dependent on the data speculative loads before reload.
12670 This is effective only with @option{-msched-br-data-spec} enabled.
12671 The default is 'enable'.
12673 @item -msched-ar-in-data-spec
12674 @itemx -mno-sched-ar-in-data-spec
12675 @opindex msched-ar-in-data-spec
12676 @opindex mno-sched-ar-in-data-spec
12677 (En/Dis)able speculative scheduling of the instructions that
12678 are dependent on the data speculative loads after reload.
12679 This is effective only with @option{-msched-ar-data-spec} enabled.
12680 The default is 'enable'.
12682 @item -msched-in-control-spec
12683 @itemx -mno-sched-in-control-spec
12684 @opindex msched-in-control-spec
12685 @opindex mno-sched-in-control-spec
12686 (En/Dis)able speculative scheduling of the instructions that
12687 are dependent on the control speculative loads.
12688 This is effective only with @option{-msched-control-spec} enabled.
12689 The default is 'enable'.
12691 @item -mno-sched-prefer-non-data-spec-insns
12692 @itemx -msched-prefer-non-data-spec-insns
12693 @opindex mno-sched-prefer-non-data-spec-insns
12694 @opindex msched-prefer-non-data-spec-insns
12695 If enabled, data speculative instructions will be chosen for schedule
12696 only if there are no other choices at the moment. This will make
12697 the use of the data speculation much more conservative.
12698 The default is 'disable'.
12700 @item -mno-sched-prefer-non-control-spec-insns
12701 @itemx -msched-prefer-non-control-spec-insns
12702 @opindex mno-sched-prefer-non-control-spec-insns
12703 @opindex msched-prefer-non-control-spec-insns
12704 If enabled, control speculative instructions will be chosen for schedule
12705 only if there are no other choices at the moment. This will make
12706 the use of the control speculation much more conservative.
12707 The default is 'disable'.
12709 @item -mno-sched-count-spec-in-critical-path
12710 @itemx -msched-count-spec-in-critical-path
12711 @opindex mno-sched-count-spec-in-critical-path
12712 @opindex msched-count-spec-in-critical-path
12713 If enabled, speculative dependencies will be considered during
12714 computation of the instructions priorities. This will make the use of the
12715 speculation a bit more conservative.
12716 The default is 'disable'.
12718 @item -msched-spec-ldc
12719 @opindex msched-spec-ldc
12720 Use a simple data speculation check. This option is on by default.
12722 @item -msched-control-spec-ldc
12723 @opindex msched-spec-ldc
12724 Use a simple check for control speculation. This option is on by default.
12726 @item -msched-stop-bits-after-every-cycle
12727 @opindex msched-stop-bits-after-every-cycle
12728 Place a stop bit after every cycle when scheduling. This option is on
12731 @item -msched-fp-mem-deps-zero-cost
12732 @opindex msched-fp-mem-deps-zero-cost
12733 Assume that floating-point stores and loads are not likely to cause a conflict
12734 when placed into the same instruction group. This option is disabled by
12737 @item -msel-sched-dont-check-control-spec
12738 @opindex msel-sched-dont-check-control-spec
12739 Generate checks for control speculation in selective scheduling.
12740 This flag is disabled by default.
12742 @item -msched-max-memory-insns=@var{max-insns}
12743 @opindex msched-max-memory-insns
12744 Limit on the number of memory insns per instruction group, giving lower
12745 priority to subsequent memory insns attempting to schedule in the same
12746 instruction group. Frequently useful to prevent cache bank conflicts.
12747 The default value is 1.
12749 @item -msched-max-memory-insns-hard-limit
12750 @opindex msched-max-memory-insns-hard-limit
12751 Disallow more than `msched-max-memory-insns' in instruction group.
12752 Otherwise, limit is `soft' meaning that we would prefer non-memory operations
12753 when limit is reached but may still schedule memory operations.
12757 @node IA-64/VMS Options
12758 @subsection IA-64/VMS Options
12760 These @samp{-m} options are defined for the IA-64/VMS implementations:
12763 @item -mvms-return-codes
12764 @opindex mvms-return-codes
12765 Return VMS condition codes from main. The default is to return POSIX
12766 style condition (e.g.@ error) codes.
12768 @item -mdebug-main=@var{prefix}
12769 @opindex mdebug-main=@var{prefix}
12770 Flag the first routine whose name starts with @var{prefix} as the main
12771 routine for the debugger.
12775 Default to 64bit memory allocation routines.
12779 @subsection LM32 Options
12780 @cindex LM32 options
12782 These @option{-m} options are defined for the Lattice Mico32 architecture:
12785 @item -mbarrel-shift-enabled
12786 @opindex mbarrel-shift-enabled
12787 Enable barrel-shift instructions.
12789 @item -mdivide-enabled
12790 @opindex mdivide-enabled
12791 Enable divide and modulus instructions.
12793 @item -mmultiply-enabled
12794 @opindex multiply-enabled
12795 Enable multiply instructions.
12797 @item -msign-extend-enabled
12798 @opindex msign-extend-enabled
12799 Enable sign extend instructions.
12801 @item -muser-enabled
12802 @opindex muser-enabled
12803 Enable user-defined instructions.
12808 @subsection M32C Options
12809 @cindex M32C options
12812 @item -mcpu=@var{name}
12814 Select the CPU for which code is generated. @var{name} may be one of
12815 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
12816 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
12817 the M32C/80 series.
12821 Specifies that the program will be run on the simulator. This causes
12822 an alternate runtime library to be linked in which supports, for
12823 example, file I/O@. You must not use this option when generating
12824 programs that will run on real hardware; you must provide your own
12825 runtime library for whatever I/O functions are needed.
12827 @item -memregs=@var{number}
12829 Specifies the number of memory-based pseudo-registers GCC will use
12830 during code generation. These pseudo-registers will be used like real
12831 registers, so there is a tradeoff between GCC's ability to fit the
12832 code into available registers, and the performance penalty of using
12833 memory instead of registers. Note that all modules in a program must
12834 be compiled with the same value for this option. Because of that, you
12835 must not use this option with the default runtime libraries gcc
12840 @node M32R/D Options
12841 @subsection M32R/D Options
12842 @cindex M32R/D options
12844 These @option{-m} options are defined for Renesas M32R/D architectures:
12849 Generate code for the M32R/2@.
12853 Generate code for the M32R/X@.
12857 Generate code for the M32R@. This is the default.
12859 @item -mmodel=small
12860 @opindex mmodel=small
12861 Assume all objects live in the lower 16MB of memory (so that their addresses
12862 can be loaded with the @code{ld24} instruction), and assume all subroutines
12863 are reachable with the @code{bl} instruction.
12864 This is the default.
12866 The addressability of a particular object can be set with the
12867 @code{model} attribute.
12869 @item -mmodel=medium
12870 @opindex mmodel=medium
12871 Assume objects may be anywhere in the 32-bit address space (the compiler
12872 will generate @code{seth/add3} instructions to load their addresses), and
12873 assume all subroutines are reachable with the @code{bl} instruction.
12875 @item -mmodel=large
12876 @opindex mmodel=large
12877 Assume objects may be anywhere in the 32-bit address space (the compiler
12878 will generate @code{seth/add3} instructions to load their addresses), and
12879 assume subroutines may not be reachable with the @code{bl} instruction
12880 (the compiler will generate the much slower @code{seth/add3/jl}
12881 instruction sequence).
12884 @opindex msdata=none
12885 Disable use of the small data area. Variables will be put into
12886 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
12887 @code{section} attribute has been specified).
12888 This is the default.
12890 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
12891 Objects may be explicitly put in the small data area with the
12892 @code{section} attribute using one of these sections.
12894 @item -msdata=sdata
12895 @opindex msdata=sdata
12896 Put small global and static data in the small data area, but do not
12897 generate special code to reference them.
12900 @opindex msdata=use
12901 Put small global and static data in the small data area, and generate
12902 special instructions to reference them.
12906 @cindex smaller data references
12907 Put global and static objects less than or equal to @var{num} bytes
12908 into the small data or bss sections instead of the normal data or bss
12909 sections. The default value of @var{num} is 8.
12910 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
12911 for this option to have any effect.
12913 All modules should be compiled with the same @option{-G @var{num}} value.
12914 Compiling with different values of @var{num} may or may not work; if it
12915 doesn't the linker will give an error message---incorrect code will not be
12920 Makes the M32R specific code in the compiler display some statistics
12921 that might help in debugging programs.
12923 @item -malign-loops
12924 @opindex malign-loops
12925 Align all loops to a 32-byte boundary.
12927 @item -mno-align-loops
12928 @opindex mno-align-loops
12929 Do not enforce a 32-byte alignment for loops. This is the default.
12931 @item -missue-rate=@var{number}
12932 @opindex missue-rate=@var{number}
12933 Issue @var{number} instructions per cycle. @var{number} can only be 1
12936 @item -mbranch-cost=@var{number}
12937 @opindex mbranch-cost=@var{number}
12938 @var{number} can only be 1 or 2. If it is 1 then branches will be
12939 preferred over conditional code, if it is 2, then the opposite will
12942 @item -mflush-trap=@var{number}
12943 @opindex mflush-trap=@var{number}
12944 Specifies the trap number to use to flush the cache. The default is
12945 12. Valid numbers are between 0 and 15 inclusive.
12947 @item -mno-flush-trap
12948 @opindex mno-flush-trap
12949 Specifies that the cache cannot be flushed by using a trap.
12951 @item -mflush-func=@var{name}
12952 @opindex mflush-func=@var{name}
12953 Specifies the name of the operating system function to call to flush
12954 the cache. The default is @emph{_flush_cache}, but a function call
12955 will only be used if a trap is not available.
12957 @item -mno-flush-func
12958 @opindex mno-flush-func
12959 Indicates that there is no OS function for flushing the cache.
12963 @node M680x0 Options
12964 @subsection M680x0 Options
12965 @cindex M680x0 options
12967 These are the @samp{-m} options defined for M680x0 and ColdFire processors.
12968 The default settings depend on which architecture was selected when
12969 the compiler was configured; the defaults for the most common choices
12973 @item -march=@var{arch}
12975 Generate code for a specific M680x0 or ColdFire instruction set
12976 architecture. Permissible values of @var{arch} for M680x0
12977 architectures are: @samp{68000}, @samp{68010}, @samp{68020},
12978 @samp{68030}, @samp{68040}, @samp{68060} and @samp{cpu32}. ColdFire
12979 architectures are selected according to Freescale's ISA classification
12980 and the permissible values are: @samp{isaa}, @samp{isaaplus},
12981 @samp{isab} and @samp{isac}.
12983 gcc defines a macro @samp{__mcf@var{arch}__} whenever it is generating
12984 code for a ColdFire target. The @var{arch} in this macro is one of the
12985 @option{-march} arguments given above.
12987 When used together, @option{-march} and @option{-mtune} select code
12988 that runs on a family of similar processors but that is optimized
12989 for a particular microarchitecture.
12991 @item -mcpu=@var{cpu}
12993 Generate code for a specific M680x0 or ColdFire processor.
12994 The M680x0 @var{cpu}s are: @samp{68000}, @samp{68010}, @samp{68020},
12995 @samp{68030}, @samp{68040}, @samp{68060}, @samp{68302}, @samp{68332}
12996 and @samp{cpu32}. The ColdFire @var{cpu}s are given by the table
12997 below, which also classifies the CPUs into families:
12999 @multitable @columnfractions 0.20 0.80
13000 @item @strong{Family} @tab @strong{@samp{-mcpu} arguments}
13001 @item @samp{51} @tab @samp{51} @samp{51ac} @samp{51cn} @samp{51em} @samp{51qe}
13002 @item @samp{5206} @tab @samp{5202} @samp{5204} @samp{5206}
13003 @item @samp{5206e} @tab @samp{5206e}
13004 @item @samp{5208} @tab @samp{5207} @samp{5208}
13005 @item @samp{5211a} @tab @samp{5210a} @samp{5211a}
13006 @item @samp{5213} @tab @samp{5211} @samp{5212} @samp{5213}
13007 @item @samp{5216} @tab @samp{5214} @samp{5216}
13008 @item @samp{52235} @tab @samp{52230} @samp{52231} @samp{52232} @samp{52233} @samp{52234} @samp{52235}
13009 @item @samp{5225} @tab @samp{5224} @samp{5225}
13010 @item @samp{52259} @tab @samp{52252} @samp{52254} @samp{52255} @samp{52256} @samp{52258} @samp{52259}
13011 @item @samp{5235} @tab @samp{5232} @samp{5233} @samp{5234} @samp{5235} @samp{523x}
13012 @item @samp{5249} @tab @samp{5249}
13013 @item @samp{5250} @tab @samp{5250}
13014 @item @samp{5271} @tab @samp{5270} @samp{5271}
13015 @item @samp{5272} @tab @samp{5272}
13016 @item @samp{5275} @tab @samp{5274} @samp{5275}
13017 @item @samp{5282} @tab @samp{5280} @samp{5281} @samp{5282} @samp{528x}
13018 @item @samp{53017} @tab @samp{53011} @samp{53012} @samp{53013} @samp{53014} @samp{53015} @samp{53016} @samp{53017}
13019 @item @samp{5307} @tab @samp{5307}
13020 @item @samp{5329} @tab @samp{5327} @samp{5328} @samp{5329} @samp{532x}
13021 @item @samp{5373} @tab @samp{5372} @samp{5373} @samp{537x}
13022 @item @samp{5407} @tab @samp{5407}
13023 @item @samp{5475} @tab @samp{5470} @samp{5471} @samp{5472} @samp{5473} @samp{5474} @samp{5475} @samp{547x} @samp{5480} @samp{5481} @samp{5482} @samp{5483} @samp{5484} @samp{5485}
13026 @option{-mcpu=@var{cpu}} overrides @option{-march=@var{arch}} if
13027 @var{arch} is compatible with @var{cpu}. Other combinations of
13028 @option{-mcpu} and @option{-march} are rejected.
13030 gcc defines the macro @samp{__mcf_cpu_@var{cpu}} when ColdFire target
13031 @var{cpu} is selected. It also defines @samp{__mcf_family_@var{family}},
13032 where the value of @var{family} is given by the table above.
13034 @item -mtune=@var{tune}
13036 Tune the code for a particular microarchitecture, within the
13037 constraints set by @option{-march} and @option{-mcpu}.
13038 The M680x0 microarchitectures are: @samp{68000}, @samp{68010},
13039 @samp{68020}, @samp{68030}, @samp{68040}, @samp{68060}
13040 and @samp{cpu32}. The ColdFire microarchitectures
13041 are: @samp{cfv1}, @samp{cfv2}, @samp{cfv3}, @samp{cfv4} and @samp{cfv4e}.
13043 You can also use @option{-mtune=68020-40} for code that needs
13044 to run relatively well on 68020, 68030 and 68040 targets.
13045 @option{-mtune=68020-60} is similar but includes 68060 targets
13046 as well. These two options select the same tuning decisions as
13047 @option{-m68020-40} and @option{-m68020-60} respectively.
13049 gcc defines the macros @samp{__mc@var{arch}} and @samp{__mc@var{arch}__}
13050 when tuning for 680x0 architecture @var{arch}. It also defines
13051 @samp{mc@var{arch}} unless either @option{-ansi} or a non-GNU @option{-std}
13052 option is used. If gcc is tuning for a range of architectures,
13053 as selected by @option{-mtune=68020-40} or @option{-mtune=68020-60},
13054 it defines the macros for every architecture in the range.
13056 gcc also defines the macro @samp{__m@var{uarch}__} when tuning for
13057 ColdFire microarchitecture @var{uarch}, where @var{uarch} is one
13058 of the arguments given above.
13064 Generate output for a 68000. This is the default
13065 when the compiler is configured for 68000-based systems.
13066 It is equivalent to @option{-march=68000}.
13068 Use this option for microcontrollers with a 68000 or EC000 core,
13069 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
13073 Generate output for a 68010. This is the default
13074 when the compiler is configured for 68010-based systems.
13075 It is equivalent to @option{-march=68010}.
13081 Generate output for a 68020. This is the default
13082 when the compiler is configured for 68020-based systems.
13083 It is equivalent to @option{-march=68020}.
13087 Generate output for a 68030. This is the default when the compiler is
13088 configured for 68030-based systems. It is equivalent to
13089 @option{-march=68030}.
13093 Generate output for a 68040. This is the default when the compiler is
13094 configured for 68040-based systems. It is equivalent to
13095 @option{-march=68040}.
13097 This option inhibits the use of 68881/68882 instructions that have to be
13098 emulated by software on the 68040. Use this option if your 68040 does not
13099 have code to emulate those instructions.
13103 Generate output for a 68060. This is the default when the compiler is
13104 configured for 68060-based systems. It is equivalent to
13105 @option{-march=68060}.
13107 This option inhibits the use of 68020 and 68881/68882 instructions that
13108 have to be emulated by software on the 68060. Use this option if your 68060
13109 does not have code to emulate those instructions.
13113 Generate output for a CPU32. This is the default
13114 when the compiler is configured for CPU32-based systems.
13115 It is equivalent to @option{-march=cpu32}.
13117 Use this option for microcontrollers with a
13118 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
13119 68336, 68340, 68341, 68349 and 68360.
13123 Generate output for a 520X ColdFire CPU@. This is the default
13124 when the compiler is configured for 520X-based systems.
13125 It is equivalent to @option{-mcpu=5206}, and is now deprecated
13126 in favor of that option.
13128 Use this option for microcontroller with a 5200 core, including
13129 the MCF5202, MCF5203, MCF5204 and MCF5206.
13133 Generate output for a 5206e ColdFire CPU@. The option is now
13134 deprecated in favor of the equivalent @option{-mcpu=5206e}.
13138 Generate output for a member of the ColdFire 528X family.
13139 The option is now deprecated in favor of the equivalent
13140 @option{-mcpu=528x}.
13144 Generate output for a ColdFire 5307 CPU@. The option is now deprecated
13145 in favor of the equivalent @option{-mcpu=5307}.
13149 Generate output for a ColdFire 5407 CPU@. The option is now deprecated
13150 in favor of the equivalent @option{-mcpu=5407}.
13154 Generate output for a ColdFire V4e family CPU (e.g.@: 547x/548x).
13155 This includes use of hardware floating point instructions.
13156 The option is equivalent to @option{-mcpu=547x}, and is now
13157 deprecated in favor of that option.
13161 Generate output for a 68040, without using any of the new instructions.
13162 This results in code which can run relatively efficiently on either a
13163 68020/68881 or a 68030 or a 68040. The generated code does use the
13164 68881 instructions that are emulated on the 68040.
13166 The option is equivalent to @option{-march=68020} @option{-mtune=68020-40}.
13170 Generate output for a 68060, without using any of the new instructions.
13171 This results in code which can run relatively efficiently on either a
13172 68020/68881 or a 68030 or a 68040. The generated code does use the
13173 68881 instructions that are emulated on the 68060.
13175 The option is equivalent to @option{-march=68020} @option{-mtune=68020-60}.
13179 @opindex mhard-float
13181 Generate floating-point instructions. This is the default for 68020
13182 and above, and for ColdFire devices that have an FPU@. It defines the
13183 macro @samp{__HAVE_68881__} on M680x0 targets and @samp{__mcffpu__}
13184 on ColdFire targets.
13187 @opindex msoft-float
13188 Do not generate floating-point instructions; use library calls instead.
13189 This is the default for 68000, 68010, and 68832 targets. It is also
13190 the default for ColdFire devices that have no FPU.
13196 Generate (do not generate) ColdFire hardware divide and remainder
13197 instructions. If @option{-march} is used without @option{-mcpu},
13198 the default is ``on'' for ColdFire architectures and ``off'' for M680x0
13199 architectures. Otherwise, the default is taken from the target CPU
13200 (either the default CPU, or the one specified by @option{-mcpu}). For
13201 example, the default is ``off'' for @option{-mcpu=5206} and ``on'' for
13202 @option{-mcpu=5206e}.
13204 gcc defines the macro @samp{__mcfhwdiv__} when this option is enabled.
13208 Consider type @code{int} to be 16 bits wide, like @code{short int}.
13209 Additionally, parameters passed on the stack are also aligned to a
13210 16-bit boundary even on targets whose API mandates promotion to 32-bit.
13214 Do not consider type @code{int} to be 16 bits wide. This is the default.
13217 @itemx -mno-bitfield
13218 @opindex mnobitfield
13219 @opindex mno-bitfield
13220 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
13221 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
13225 Do use the bit-field instructions. The @option{-m68020} option implies
13226 @option{-mbitfield}. This is the default if you use a configuration
13227 designed for a 68020.
13231 Use a different function-calling convention, in which functions
13232 that take a fixed number of arguments return with the @code{rtd}
13233 instruction, which pops their arguments while returning. This
13234 saves one instruction in the caller since there is no need to pop
13235 the arguments there.
13237 This calling convention is incompatible with the one normally
13238 used on Unix, so you cannot use it if you need to call libraries
13239 compiled with the Unix compiler.
13241 Also, you must provide function prototypes for all functions that
13242 take variable numbers of arguments (including @code{printf});
13243 otherwise incorrect code will be generated for calls to those
13246 In addition, seriously incorrect code will result if you call a
13247 function with too many arguments. (Normally, extra arguments are
13248 harmlessly ignored.)
13250 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
13251 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
13255 Do not use the calling conventions selected by @option{-mrtd}.
13256 This is the default.
13259 @itemx -mno-align-int
13260 @opindex malign-int
13261 @opindex mno-align-int
13262 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
13263 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
13264 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
13265 Aligning variables on 32-bit boundaries produces code that runs somewhat
13266 faster on processors with 32-bit busses at the expense of more memory.
13268 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
13269 align structures containing the above types differently than
13270 most published application binary interface specifications for the m68k.
13274 Use the pc-relative addressing mode of the 68000 directly, instead of
13275 using a global offset table. At present, this option implies @option{-fpic},
13276 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
13277 not presently supported with @option{-mpcrel}, though this could be supported for
13278 68020 and higher processors.
13280 @item -mno-strict-align
13281 @itemx -mstrict-align
13282 @opindex mno-strict-align
13283 @opindex mstrict-align
13284 Do not (do) assume that unaligned memory references will be handled by
13288 Generate code that allows the data segment to be located in a different
13289 area of memory from the text segment. This allows for execute in place in
13290 an environment without virtual memory management. This option implies
13293 @item -mno-sep-data
13294 Generate code that assumes that the data segment follows the text segment.
13295 This is the default.
13297 @item -mid-shared-library
13298 Generate code that supports shared libraries via the library ID method.
13299 This allows for execute in place and shared libraries in an environment
13300 without virtual memory management. This option implies @option{-fPIC}.
13302 @item -mno-id-shared-library
13303 Generate code that doesn't assume ID based shared libraries are being used.
13304 This is the default.
13306 @item -mshared-library-id=n
13307 Specified the identification number of the ID based shared library being
13308 compiled. Specifying a value of 0 will generate more compact code, specifying
13309 other values will force the allocation of that number to the current
13310 library but is no more space or time efficient than omitting this option.
13316 When generating position-independent code for ColdFire, generate code
13317 that works if the GOT has more than 8192 entries. This code is
13318 larger and slower than code generated without this option. On M680x0
13319 processors, this option is not needed; @option{-fPIC} suffices.
13321 GCC normally uses a single instruction to load values from the GOT@.
13322 While this is relatively efficient, it only works if the GOT
13323 is smaller than about 64k. Anything larger causes the linker
13324 to report an error such as:
13326 @cindex relocation truncated to fit (ColdFire)
13328 relocation truncated to fit: R_68K_GOT16O foobar
13331 If this happens, you should recompile your code with @option{-mxgot}.
13332 It should then work with very large GOTs. However, code generated with
13333 @option{-mxgot} is less efficient, since it takes 4 instructions to fetch
13334 the value of a global symbol.
13336 Note that some linkers, including newer versions of the GNU linker,
13337 can create multiple GOTs and sort GOT entries. If you have such a linker,
13338 you should only need to use @option{-mxgot} when compiling a single
13339 object file that accesses more than 8192 GOT entries. Very few do.
13341 These options have no effect unless GCC is generating
13342 position-independent code.
13346 @node M68hc1x Options
13347 @subsection M68hc1x Options
13348 @cindex M68hc1x options
13350 These are the @samp{-m} options defined for the 68hc11 and 68hc12
13351 microcontrollers. The default values for these options depends on
13352 which style of microcontroller was selected when the compiler was configured;
13353 the defaults for the most common choices are given below.
13360 Generate output for a 68HC11. This is the default
13361 when the compiler is configured for 68HC11-based systems.
13367 Generate output for a 68HC12. This is the default
13368 when the compiler is configured for 68HC12-based systems.
13374 Generate output for a 68HCS12.
13376 @item -mauto-incdec
13377 @opindex mauto-incdec
13378 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
13385 Enable the use of 68HC12 min and max instructions.
13388 @itemx -mno-long-calls
13389 @opindex mlong-calls
13390 @opindex mno-long-calls
13391 Treat all calls as being far away (near). If calls are assumed to be
13392 far away, the compiler will use the @code{call} instruction to
13393 call a function and the @code{rtc} instruction for returning.
13397 Consider type @code{int} to be 16 bits wide, like @code{short int}.
13399 @item -msoft-reg-count=@var{count}
13400 @opindex msoft-reg-count
13401 Specify the number of pseudo-soft registers which are used for the
13402 code generation. The maximum number is 32. Using more pseudo-soft
13403 register may or may not result in better code depending on the program.
13404 The default is 4 for 68HC11 and 2 for 68HC12.
13408 @node MCore Options
13409 @subsection MCore Options
13410 @cindex MCore options
13412 These are the @samp{-m} options defined for the Motorola M*Core
13418 @itemx -mno-hardlit
13420 @opindex mno-hardlit
13421 Inline constants into the code stream if it can be done in two
13422 instructions or less.
13428 Use the divide instruction. (Enabled by default).
13430 @item -mrelax-immediate
13431 @itemx -mno-relax-immediate
13432 @opindex mrelax-immediate
13433 @opindex mno-relax-immediate
13434 Allow arbitrary sized immediates in bit operations.
13436 @item -mwide-bitfields
13437 @itemx -mno-wide-bitfields
13438 @opindex mwide-bitfields
13439 @opindex mno-wide-bitfields
13440 Always treat bit-fields as int-sized.
13442 @item -m4byte-functions
13443 @itemx -mno-4byte-functions
13444 @opindex m4byte-functions
13445 @opindex mno-4byte-functions
13446 Force all functions to be aligned to a four byte boundary.
13448 @item -mcallgraph-data
13449 @itemx -mno-callgraph-data
13450 @opindex mcallgraph-data
13451 @opindex mno-callgraph-data
13452 Emit callgraph information.
13455 @itemx -mno-slow-bytes
13456 @opindex mslow-bytes
13457 @opindex mno-slow-bytes
13458 Prefer word access when reading byte quantities.
13460 @item -mlittle-endian
13461 @itemx -mbig-endian
13462 @opindex mlittle-endian
13463 @opindex mbig-endian
13464 Generate code for a little endian target.
13470 Generate code for the 210 processor.
13474 Assume that run-time support has been provided and so omit the
13475 simulator library (@file{libsim.a)} from the linker command line.
13477 @item -mstack-increment=@var{size}
13478 @opindex mstack-increment
13479 Set the maximum amount for a single stack increment operation. Large
13480 values can increase the speed of programs which contain functions
13481 that need a large amount of stack space, but they can also trigger a
13482 segmentation fault if the stack is extended too much. The default
13488 @subsection MeP Options
13489 @cindex MeP options
13495 Enables the @code{abs} instruction, which is the absolute difference
13496 between two registers.
13500 Enables all the optional instructions - average, multiply, divide, bit
13501 operations, leading zero, absolute difference, min/max, clip, and
13507 Enables the @code{ave} instruction, which computes the average of two
13510 @item -mbased=@var{n}
13512 Variables of size @var{n} bytes or smaller will be placed in the
13513 @code{.based} section by default. Based variables use the @code{$tp}
13514 register as a base register, and there is a 128 byte limit to the
13515 @code{.based} section.
13519 Enables the bit operation instructions - bit test (@code{btstm}), set
13520 (@code{bsetm}), clear (@code{bclrm}), invert (@code{bnotm}), and
13521 test-and-set (@code{tas}).
13523 @item -mc=@var{name}
13525 Selects which section constant data will be placed in. @var{name} may
13526 be @code{tiny}, @code{near}, or @code{far}.
13530 Enables the @code{clip} instruction. Note that @code{-mclip} is not
13531 useful unless you also provide @code{-mminmax}.
13533 @item -mconfig=@var{name}
13535 Selects one of the build-in core configurations. Each MeP chip has
13536 one or more modules in it; each module has a core CPU and a variety of
13537 coprocessors, optional instructions, and peripherals. The
13538 @code{MeP-Integrator} tool, not part of GCC, provides these
13539 configurations through this option; using this option is the same as
13540 using all the corresponding command line options. The default
13541 configuration is @code{default}.
13545 Enables the coprocessor instructions. By default, this is a 32-bit
13546 coprocessor. Note that the coprocessor is normally enabled via the
13547 @code{-mconfig=} option.
13551 Enables the 32-bit coprocessor's instructions.
13555 Enables the 64-bit coprocessor's instructions.
13559 Enables IVC2 scheduling. IVC2 is a 64-bit VLIW coprocessor.
13563 Causes constant variables to be placed in the @code{.near} section.
13567 Enables the @code{div} and @code{divu} instructions.
13571 Generate big-endian code.
13575 Generate little-endian code.
13577 @item -mio-volatile
13578 @opindex mio-volatile
13579 Tells the compiler that any variable marked with the @code{io}
13580 attribute is to be considered volatile.
13584 Causes variables to be assigned to the @code{.far} section by default.
13588 Enables the @code{leadz} (leading zero) instruction.
13592 Causes variables to be assigned to the @code{.near} section by default.
13596 Enables the @code{min} and @code{max} instructions.
13600 Enables the multiplication and multiply-accumulate instructions.
13604 Disables all the optional instructions enabled by @code{-mall-opts}.
13608 Enables the @code{repeat} and @code{erepeat} instructions, used for
13609 low-overhead looping.
13613 Causes all variables to default to the @code{.tiny} section. Note
13614 that there is a 65536 byte limit to this section. Accesses to these
13615 variables use the @code{%gp} base register.
13619 Enables the saturation instructions. Note that the compiler does not
13620 currently generate these itself, but this option is included for
13621 compatibility with other tools, like @code{as}.
13625 Link the SDRAM-based runtime instead of the default ROM-based runtime.
13629 Link the simulator runtime libraries.
13633 Link the simulator runtime libraries, excluding built-in support
13634 for reset and exception vectors and tables.
13638 Causes all functions to default to the @code{.far} section. Without
13639 this option, functions default to the @code{.near} section.
13641 @item -mtiny=@var{n}
13643 Variables that are @var{n} bytes or smaller will be allocated to the
13644 @code{.tiny} section. These variables use the @code{$gp} base
13645 register. The default for this option is 4, but note that there's a
13646 65536 byte limit to the @code{.tiny} section.
13651 @subsection MIPS Options
13652 @cindex MIPS options
13658 Generate big-endian code.
13662 Generate little-endian code. This is the default for @samp{mips*el-*-*}
13665 @item -march=@var{arch}
13667 Generate code that will run on @var{arch}, which can be the name of a
13668 generic MIPS ISA, or the name of a particular processor.
13670 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
13671 @samp{mips32}, @samp{mips32r2}, @samp{mips64} and @samp{mips64r2}.
13672 The processor names are:
13673 @samp{4kc}, @samp{4km}, @samp{4kp}, @samp{4ksc},
13674 @samp{4kec}, @samp{4kem}, @samp{4kep}, @samp{4ksd},
13675 @samp{5kc}, @samp{5kf},
13677 @samp{24kc}, @samp{24kf2_1}, @samp{24kf1_1},
13678 @samp{24kec}, @samp{24kef2_1}, @samp{24kef1_1},
13679 @samp{34kc}, @samp{34kf2_1}, @samp{34kf1_1},
13680 @samp{74kc}, @samp{74kf2_1}, @samp{74kf1_1}, @samp{74kf3_2},
13681 @samp{1004kc}, @samp{1004kf2_1}, @samp{1004kf1_1},
13682 @samp{loongson2e}, @samp{loongson2f},
13686 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
13687 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000},
13688 @samp{rm7000}, @samp{rm9000},
13689 @samp{r10000}, @samp{r12000}, @samp{r14000}, @samp{r16000},
13692 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
13693 @samp{vr5000}, @samp{vr5400}, @samp{vr5500}
13695 The special value @samp{from-abi} selects the
13696 most compatible architecture for the selected ABI (that is,
13697 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
13699 Native Linux/GNU toolchains also support the value @samp{native},
13700 which selects the best architecture option for the host processor.
13701 @option{-march=native} has no effect if GCC does not recognize
13704 In processor names, a final @samp{000} can be abbreviated as @samp{k}
13705 (for example, @samp{-march=r2k}). Prefixes are optional, and
13706 @samp{vr} may be written @samp{r}.
13708 Names of the form @samp{@var{n}f2_1} refer to processors with
13709 FPUs clocked at half the rate of the core, names of the form
13710 @samp{@var{n}f1_1} refer to processors with FPUs clocked at the same
13711 rate as the core, and names of the form @samp{@var{n}f3_2} refer to
13712 processors with FPUs clocked a ratio of 3:2 with respect to the core.
13713 For compatibility reasons, @samp{@var{n}f} is accepted as a synonym
13714 for @samp{@var{n}f2_1} while @samp{@var{n}x} and @samp{@var{b}fx} are
13715 accepted as synonyms for @samp{@var{n}f1_1}.
13717 GCC defines two macros based on the value of this option. The first
13718 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
13719 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
13720 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
13721 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
13722 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
13724 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
13725 above. In other words, it will have the full prefix and will not
13726 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
13727 the macro names the resolved architecture (either @samp{"mips1"} or
13728 @samp{"mips3"}). It names the default architecture when no
13729 @option{-march} option is given.
13731 @item -mtune=@var{arch}
13733 Optimize for @var{arch}. Among other things, this option controls
13734 the way instructions are scheduled, and the perceived cost of arithmetic
13735 operations. The list of @var{arch} values is the same as for
13738 When this option is not used, GCC will optimize for the processor
13739 specified by @option{-march}. By using @option{-march} and
13740 @option{-mtune} together, it is possible to generate code that will
13741 run on a family of processors, but optimize the code for one
13742 particular member of that family.
13744 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
13745 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
13746 @samp{-march} ones described above.
13750 Equivalent to @samp{-march=mips1}.
13754 Equivalent to @samp{-march=mips2}.
13758 Equivalent to @samp{-march=mips3}.
13762 Equivalent to @samp{-march=mips4}.
13766 Equivalent to @samp{-march=mips32}.
13770 Equivalent to @samp{-march=mips32r2}.
13774 Equivalent to @samp{-march=mips64}.
13778 Equivalent to @samp{-march=mips64r2}.
13783 @opindex mno-mips16
13784 Generate (do not generate) MIPS16 code. If GCC is targetting a
13785 MIPS32 or MIPS64 architecture, it will make use of the MIPS16e ASE@.
13787 MIPS16 code generation can also be controlled on a per-function basis
13788 by means of @code{mips16} and @code{nomips16} attributes.
13789 @xref{Function Attributes}, for more information.
13791 @item -mflip-mips16
13792 @opindex mflip-mips16
13793 Generate MIPS16 code on alternating functions. This option is provided
13794 for regression testing of mixed MIPS16/non-MIPS16 code generation, and is
13795 not intended for ordinary use in compiling user code.
13797 @item -minterlink-mips16
13798 @itemx -mno-interlink-mips16
13799 @opindex minterlink-mips16
13800 @opindex mno-interlink-mips16
13801 Require (do not require) that non-MIPS16 code be link-compatible with
13804 For example, non-MIPS16 code cannot jump directly to MIPS16 code;
13805 it must either use a call or an indirect jump. @option{-minterlink-mips16}
13806 therefore disables direct jumps unless GCC knows that the target of the
13807 jump is not MIPS16.
13819 Generate code for the given ABI@.
13821 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
13822 generates 64-bit code when you select a 64-bit architecture, but you
13823 can use @option{-mgp32} to get 32-bit code instead.
13825 For information about the O64 ABI, see
13826 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
13828 GCC supports a variant of the o32 ABI in which floating-point registers
13829 are 64 rather than 32 bits wide. You can select this combination with
13830 @option{-mabi=32} @option{-mfp64}. This ABI relies on the @samp{mthc1}
13831 and @samp{mfhc1} instructions and is therefore only supported for
13832 MIPS32R2 processors.
13834 The register assignments for arguments and return values remain the
13835 same, but each scalar value is passed in a single 64-bit register
13836 rather than a pair of 32-bit registers. For example, scalar
13837 floating-point values are returned in @samp{$f0} only, not a
13838 @samp{$f0}/@samp{$f1} pair. The set of call-saved registers also
13839 remains the same, but all 64 bits are saved.
13842 @itemx -mno-abicalls
13844 @opindex mno-abicalls
13845 Generate (do not generate) code that is suitable for SVR4-style
13846 dynamic objects. @option{-mabicalls} is the default for SVR4-based
13851 Generate (do not generate) code that is fully position-independent,
13852 and that can therefore be linked into shared libraries. This option
13853 only affects @option{-mabicalls}.
13855 All @option{-mabicalls} code has traditionally been position-independent,
13856 regardless of options like @option{-fPIC} and @option{-fpic}. However,
13857 as an extension, the GNU toolchain allows executables to use absolute
13858 accesses for locally-binding symbols. It can also use shorter GP
13859 initialization sequences and generate direct calls to locally-defined
13860 functions. This mode is selected by @option{-mno-shared}.
13862 @option{-mno-shared} depends on binutils 2.16 or higher and generates
13863 objects that can only be linked by the GNU linker. However, the option
13864 does not affect the ABI of the final executable; it only affects the ABI
13865 of relocatable objects. Using @option{-mno-shared} will generally make
13866 executables both smaller and quicker.
13868 @option{-mshared} is the default.
13874 Assume (do not assume) that the static and dynamic linkers
13875 support PLTs and copy relocations. This option only affects
13876 @samp{-mno-shared -mabicalls}. For the n64 ABI, this option
13877 has no effect without @samp{-msym32}.
13879 You can make @option{-mplt} the default by configuring
13880 GCC with @option{--with-mips-plt}. The default is
13881 @option{-mno-plt} otherwise.
13887 Lift (do not lift) the usual restrictions on the size of the global
13890 GCC normally uses a single instruction to load values from the GOT@.
13891 While this is relatively efficient, it will only work if the GOT
13892 is smaller than about 64k. Anything larger will cause the linker
13893 to report an error such as:
13895 @cindex relocation truncated to fit (MIPS)
13897 relocation truncated to fit: R_MIPS_GOT16 foobar
13900 If this happens, you should recompile your code with @option{-mxgot}.
13901 It should then work with very large GOTs, although it will also be
13902 less efficient, since it will take three instructions to fetch the
13903 value of a global symbol.
13905 Note that some linkers can create multiple GOTs. If you have such a
13906 linker, you should only need to use @option{-mxgot} when a single object
13907 file accesses more than 64k's worth of GOT entries. Very few do.
13909 These options have no effect unless GCC is generating position
13914 Assume that general-purpose registers are 32 bits wide.
13918 Assume that general-purpose registers are 64 bits wide.
13922 Assume that floating-point registers are 32 bits wide.
13926 Assume that floating-point registers are 64 bits wide.
13929 @opindex mhard-float
13930 Use floating-point coprocessor instructions.
13933 @opindex msoft-float
13934 Do not use floating-point coprocessor instructions. Implement
13935 floating-point calculations using library calls instead.
13937 @item -msingle-float
13938 @opindex msingle-float
13939 Assume that the floating-point coprocessor only supports single-precision
13942 @item -mdouble-float
13943 @opindex mdouble-float
13944 Assume that the floating-point coprocessor supports double-precision
13945 operations. This is the default.
13951 Use (do not use) @samp{ll}, @samp{sc}, and @samp{sync} instructions to
13952 implement atomic memory built-in functions. When neither option is
13953 specified, GCC will use the instructions if the target architecture
13956 @option{-mllsc} is useful if the runtime environment can emulate the
13957 instructions and @option{-mno-llsc} can be useful when compiling for
13958 nonstandard ISAs. You can make either option the default by
13959 configuring GCC with @option{--with-llsc} and @option{--without-llsc}
13960 respectively. @option{--with-llsc} is the default for some
13961 configurations; see the installation documentation for details.
13967 Use (do not use) revision 1 of the MIPS DSP ASE@.
13968 @xref{MIPS DSP Built-in Functions}. This option defines the
13969 preprocessor macro @samp{__mips_dsp}. It also defines
13970 @samp{__mips_dsp_rev} to 1.
13976 Use (do not use) revision 2 of the MIPS DSP ASE@.
13977 @xref{MIPS DSP Built-in Functions}. This option defines the
13978 preprocessor macros @samp{__mips_dsp} and @samp{__mips_dspr2}.
13979 It also defines @samp{__mips_dsp_rev} to 2.
13982 @itemx -mno-smartmips
13983 @opindex msmartmips
13984 @opindex mno-smartmips
13985 Use (do not use) the MIPS SmartMIPS ASE.
13987 @item -mpaired-single
13988 @itemx -mno-paired-single
13989 @opindex mpaired-single
13990 @opindex mno-paired-single
13991 Use (do not use) paired-single floating-point instructions.
13992 @xref{MIPS Paired-Single Support}. This option requires
13993 hardware floating-point support to be enabled.
13999 Use (do not use) MIPS Digital Media Extension instructions.
14000 This option can only be used when generating 64-bit code and requires
14001 hardware floating-point support to be enabled.
14006 @opindex mno-mips3d
14007 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
14008 The option @option{-mips3d} implies @option{-mpaired-single}.
14014 Use (do not use) MT Multithreading instructions.
14018 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
14019 an explanation of the default and the way that the pointer size is
14024 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
14026 The default size of @code{int}s, @code{long}s and pointers depends on
14027 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
14028 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
14029 32-bit @code{long}s. Pointers are the same size as @code{long}s,
14030 or the same size as integer registers, whichever is smaller.
14036 Assume (do not assume) that all symbols have 32-bit values, regardless
14037 of the selected ABI@. This option is useful in combination with
14038 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
14039 to generate shorter and faster references to symbolic addresses.
14043 Put definitions of externally-visible data in a small data section
14044 if that data is no bigger than @var{num} bytes. GCC can then access
14045 the data more efficiently; see @option{-mgpopt} for details.
14047 The default @option{-G} option depends on the configuration.
14049 @item -mlocal-sdata
14050 @itemx -mno-local-sdata
14051 @opindex mlocal-sdata
14052 @opindex mno-local-sdata
14053 Extend (do not extend) the @option{-G} behavior to local data too,
14054 such as to static variables in C@. @option{-mlocal-sdata} is the
14055 default for all configurations.
14057 If the linker complains that an application is using too much small data,
14058 you might want to try rebuilding the less performance-critical parts with
14059 @option{-mno-local-sdata}. You might also want to build large
14060 libraries with @option{-mno-local-sdata}, so that the libraries leave
14061 more room for the main program.
14063 @item -mextern-sdata
14064 @itemx -mno-extern-sdata
14065 @opindex mextern-sdata
14066 @opindex mno-extern-sdata
14067 Assume (do not assume) that externally-defined data will be in
14068 a small data section if that data is within the @option{-G} limit.
14069 @option{-mextern-sdata} is the default for all configurations.
14071 If you compile a module @var{Mod} with @option{-mextern-sdata} @option{-G
14072 @var{num}} @option{-mgpopt}, and @var{Mod} references a variable @var{Var}
14073 that is no bigger than @var{num} bytes, you must make sure that @var{Var}
14074 is placed in a small data section. If @var{Var} is defined by another
14075 module, you must either compile that module with a high-enough
14076 @option{-G} setting or attach a @code{section} attribute to @var{Var}'s
14077 definition. If @var{Var} is common, you must link the application
14078 with a high-enough @option{-G} setting.
14080 The easiest way of satisfying these restrictions is to compile
14081 and link every module with the same @option{-G} option. However,
14082 you may wish to build a library that supports several different
14083 small data limits. You can do this by compiling the library with
14084 the highest supported @option{-G} setting and additionally using
14085 @option{-mno-extern-sdata} to stop the library from making assumptions
14086 about externally-defined data.
14092 Use (do not use) GP-relative accesses for symbols that are known to be
14093 in a small data section; see @option{-G}, @option{-mlocal-sdata} and
14094 @option{-mextern-sdata}. @option{-mgpopt} is the default for all
14097 @option{-mno-gpopt} is useful for cases where the @code{$gp} register
14098 might not hold the value of @code{_gp}. For example, if the code is
14099 part of a library that might be used in a boot monitor, programs that
14100 call boot monitor routines will pass an unknown value in @code{$gp}.
14101 (In such situations, the boot monitor itself would usually be compiled
14102 with @option{-G0}.)
14104 @option{-mno-gpopt} implies @option{-mno-local-sdata} and
14105 @option{-mno-extern-sdata}.
14107 @item -membedded-data
14108 @itemx -mno-embedded-data
14109 @opindex membedded-data
14110 @opindex mno-embedded-data
14111 Allocate variables to the read-only data section first if possible, then
14112 next in the small data section if possible, otherwise in data. This gives
14113 slightly slower code than the default, but reduces the amount of RAM required
14114 when executing, and thus may be preferred for some embedded systems.
14116 @item -muninit-const-in-rodata
14117 @itemx -mno-uninit-const-in-rodata
14118 @opindex muninit-const-in-rodata
14119 @opindex mno-uninit-const-in-rodata
14120 Put uninitialized @code{const} variables in the read-only data section.
14121 This option is only meaningful in conjunction with @option{-membedded-data}.
14123 @item -mcode-readable=@var{setting}
14124 @opindex mcode-readable
14125 Specify whether GCC may generate code that reads from executable sections.
14126 There are three possible settings:
14129 @item -mcode-readable=yes
14130 Instructions may freely access executable sections. This is the
14133 @item -mcode-readable=pcrel
14134 MIPS16 PC-relative load instructions can access executable sections,
14135 but other instructions must not do so. This option is useful on 4KSc
14136 and 4KSd processors when the code TLBs have the Read Inhibit bit set.
14137 It is also useful on processors that can be configured to have a dual
14138 instruction/data SRAM interface and that, like the M4K, automatically
14139 redirect PC-relative loads to the instruction RAM.
14141 @item -mcode-readable=no
14142 Instructions must not access executable sections. This option can be
14143 useful on targets that are configured to have a dual instruction/data
14144 SRAM interface but that (unlike the M4K) do not automatically redirect
14145 PC-relative loads to the instruction RAM.
14148 @item -msplit-addresses
14149 @itemx -mno-split-addresses
14150 @opindex msplit-addresses
14151 @opindex mno-split-addresses
14152 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
14153 relocation operators. This option has been superseded by
14154 @option{-mexplicit-relocs} but is retained for backwards compatibility.
14156 @item -mexplicit-relocs
14157 @itemx -mno-explicit-relocs
14158 @opindex mexplicit-relocs
14159 @opindex mno-explicit-relocs
14160 Use (do not use) assembler relocation operators when dealing with symbolic
14161 addresses. The alternative, selected by @option{-mno-explicit-relocs},
14162 is to use assembler macros instead.
14164 @option{-mexplicit-relocs} is the default if GCC was configured
14165 to use an assembler that supports relocation operators.
14167 @item -mcheck-zero-division
14168 @itemx -mno-check-zero-division
14169 @opindex mcheck-zero-division
14170 @opindex mno-check-zero-division
14171 Trap (do not trap) on integer division by zero.
14173 The default is @option{-mcheck-zero-division}.
14175 @item -mdivide-traps
14176 @itemx -mdivide-breaks
14177 @opindex mdivide-traps
14178 @opindex mdivide-breaks
14179 MIPS systems check for division by zero by generating either a
14180 conditional trap or a break instruction. Using traps results in
14181 smaller code, but is only supported on MIPS II and later. Also, some
14182 versions of the Linux kernel have a bug that prevents trap from
14183 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
14184 allow conditional traps on architectures that support them and
14185 @option{-mdivide-breaks} to force the use of breaks.
14187 The default is usually @option{-mdivide-traps}, but this can be
14188 overridden at configure time using @option{--with-divide=breaks}.
14189 Divide-by-zero checks can be completely disabled using
14190 @option{-mno-check-zero-division}.
14195 @opindex mno-memcpy
14196 Force (do not force) the use of @code{memcpy()} for non-trivial block
14197 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
14198 most constant-sized copies.
14201 @itemx -mno-long-calls
14202 @opindex mlong-calls
14203 @opindex mno-long-calls
14204 Disable (do not disable) use of the @code{jal} instruction. Calling
14205 functions using @code{jal} is more efficient but requires the caller
14206 and callee to be in the same 256 megabyte segment.
14208 This option has no effect on abicalls code. The default is
14209 @option{-mno-long-calls}.
14215 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
14216 instructions, as provided by the R4650 ISA@.
14219 @itemx -mno-fused-madd
14220 @opindex mfused-madd
14221 @opindex mno-fused-madd
14222 Enable (disable) use of the floating point multiply-accumulate
14223 instructions, when they are available. The default is
14224 @option{-mfused-madd}.
14226 When multiply-accumulate instructions are used, the intermediate
14227 product is calculated to infinite precision and is not subject to
14228 the FCSR Flush to Zero bit. This may be undesirable in some
14233 Tell the MIPS assembler to not run its preprocessor over user
14234 assembler files (with a @samp{.s} suffix) when assembling them.
14237 @itemx -mno-fix-r4000
14238 @opindex mfix-r4000
14239 @opindex mno-fix-r4000
14240 Work around certain R4000 CPU errata:
14243 A double-word or a variable shift may give an incorrect result if executed
14244 immediately after starting an integer division.
14246 A double-word or a variable shift may give an incorrect result if executed
14247 while an integer multiplication is in progress.
14249 An integer division may give an incorrect result if started in a delay slot
14250 of a taken branch or a jump.
14254 @itemx -mno-fix-r4400
14255 @opindex mfix-r4400
14256 @opindex mno-fix-r4400
14257 Work around certain R4400 CPU errata:
14260 A double-word or a variable shift may give an incorrect result if executed
14261 immediately after starting an integer division.
14265 @itemx -mno-fix-r10000
14266 @opindex mfix-r10000
14267 @opindex mno-fix-r10000
14268 Work around certain R10000 errata:
14271 @code{ll}/@code{sc} sequences may not behave atomically on revisions
14272 prior to 3.0. They may deadlock on revisions 2.6 and earlier.
14275 This option can only be used if the target architecture supports
14276 branch-likely instructions. @option{-mfix-r10000} is the default when
14277 @option{-march=r10000} is used; @option{-mno-fix-r10000} is the default
14281 @itemx -mno-fix-vr4120
14282 @opindex mfix-vr4120
14283 Work around certain VR4120 errata:
14286 @code{dmultu} does not always produce the correct result.
14288 @code{div} and @code{ddiv} do not always produce the correct result if one
14289 of the operands is negative.
14291 The workarounds for the division errata rely on special functions in
14292 @file{libgcc.a}. At present, these functions are only provided by
14293 the @code{mips64vr*-elf} configurations.
14295 Other VR4120 errata require a nop to be inserted between certain pairs of
14296 instructions. These errata are handled by the assembler, not by GCC itself.
14299 @opindex mfix-vr4130
14300 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
14301 workarounds are implemented by the assembler rather than by GCC,
14302 although GCC will avoid using @code{mflo} and @code{mfhi} if the
14303 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
14304 instructions are available instead.
14307 @itemx -mno-fix-sb1
14309 Work around certain SB-1 CPU core errata.
14310 (This flag currently works around the SB-1 revision 2
14311 ``F1'' and ``F2'' floating point errata.)
14313 @item -mr10k-cache-barrier=@var{setting}
14314 @opindex mr10k-cache-barrier
14315 Specify whether GCC should insert cache barriers to avoid the
14316 side-effects of speculation on R10K processors.
14318 In common with many processors, the R10K tries to predict the outcome
14319 of a conditional branch and speculatively executes instructions from
14320 the ``taken'' branch. It later aborts these instructions if the
14321 predicted outcome was wrong. However, on the R10K, even aborted
14322 instructions can have side effects.
14324 This problem only affects kernel stores and, depending on the system,
14325 kernel loads. As an example, a speculatively-executed store may load
14326 the target memory into cache and mark the cache line as dirty, even if
14327 the store itself is later aborted. If a DMA operation writes to the
14328 same area of memory before the ``dirty'' line is flushed, the cached
14329 data will overwrite the DMA-ed data. See the R10K processor manual
14330 for a full description, including other potential problems.
14332 One workaround is to insert cache barrier instructions before every memory
14333 access that might be speculatively executed and that might have side
14334 effects even if aborted. @option{-mr10k-cache-barrier=@var{setting}}
14335 controls GCC's implementation of this workaround. It assumes that
14336 aborted accesses to any byte in the following regions will not have
14341 the memory occupied by the current function's stack frame;
14344 the memory occupied by an incoming stack argument;
14347 the memory occupied by an object with a link-time-constant address.
14350 It is the kernel's responsibility to ensure that speculative
14351 accesses to these regions are indeed safe.
14353 If the input program contains a function declaration such as:
14359 then the implementation of @code{foo} must allow @code{j foo} and
14360 @code{jal foo} to be executed speculatively. GCC honors this
14361 restriction for functions it compiles itself. It expects non-GCC
14362 functions (such as hand-written assembly code) to do the same.
14364 The option has three forms:
14367 @item -mr10k-cache-barrier=load-store
14368 Insert a cache barrier before a load or store that might be
14369 speculatively executed and that might have side effects even
14372 @item -mr10k-cache-barrier=store
14373 Insert a cache barrier before a store that might be speculatively
14374 executed and that might have side effects even if aborted.
14376 @item -mr10k-cache-barrier=none
14377 Disable the insertion of cache barriers. This is the default setting.
14380 @item -mflush-func=@var{func}
14381 @itemx -mno-flush-func
14382 @opindex mflush-func
14383 Specifies the function to call to flush the I and D caches, or to not
14384 call any such function. If called, the function must take the same
14385 arguments as the common @code{_flush_func()}, that is, the address of the
14386 memory range for which the cache is being flushed, the size of the
14387 memory range, and the number 3 (to flush both caches). The default
14388 depends on the target GCC was configured for, but commonly is either
14389 @samp{_flush_func} or @samp{__cpu_flush}.
14391 @item mbranch-cost=@var{num}
14392 @opindex mbranch-cost
14393 Set the cost of branches to roughly @var{num} ``simple'' instructions.
14394 This cost is only a heuristic and is not guaranteed to produce
14395 consistent results across releases. A zero cost redundantly selects
14396 the default, which is based on the @option{-mtune} setting.
14398 @item -mbranch-likely
14399 @itemx -mno-branch-likely
14400 @opindex mbranch-likely
14401 @opindex mno-branch-likely
14402 Enable or disable use of Branch Likely instructions, regardless of the
14403 default for the selected architecture. By default, Branch Likely
14404 instructions may be generated if they are supported by the selected
14405 architecture. An exception is for the MIPS32 and MIPS64 architectures
14406 and processors which implement those architectures; for those, Branch
14407 Likely instructions will not be generated by default because the MIPS32
14408 and MIPS64 architectures specifically deprecate their use.
14410 @item -mfp-exceptions
14411 @itemx -mno-fp-exceptions
14412 @opindex mfp-exceptions
14413 Specifies whether FP exceptions are enabled. This affects how we schedule
14414 FP instructions for some processors. The default is that FP exceptions are
14417 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
14418 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
14421 @item -mvr4130-align
14422 @itemx -mno-vr4130-align
14423 @opindex mvr4130-align
14424 The VR4130 pipeline is two-way superscalar, but can only issue two
14425 instructions together if the first one is 8-byte aligned. When this
14426 option is enabled, GCC will align pairs of instructions that it
14427 thinks should execute in parallel.
14429 This option only has an effect when optimizing for the VR4130.
14430 It normally makes code faster, but at the expense of making it bigger.
14431 It is enabled by default at optimization level @option{-O3}.
14436 Enable (disable) generation of @code{synci} instructions on
14437 architectures that support it. The @code{synci} instructions (if
14438 enabled) will be generated when @code{__builtin___clear_cache()} is
14441 This option defaults to @code{-mno-synci}, but the default can be
14442 overridden by configuring with @code{--with-synci}.
14444 When compiling code for single processor systems, it is generally safe
14445 to use @code{synci}. However, on many multi-core (SMP) systems, it
14446 will not invalidate the instruction caches on all cores and may lead
14447 to undefined behavior.
14449 @item -mrelax-pic-calls
14450 @itemx -mno-relax-pic-calls
14451 @opindex mrelax-pic-calls
14452 Try to turn PIC calls that are normally dispatched via register
14453 @code{$25} into direct calls. This is only possible if the linker can
14454 resolve the destination at link-time and if the destination is within
14455 range for a direct call.
14457 @option{-mrelax-pic-calls} is the default if GCC was configured to use
14458 an assembler and a linker that supports the @code{.reloc} assembly
14459 directive and @code{-mexplicit-relocs} is in effect. With
14460 @code{-mno-explicit-relocs}, this optimization can be performed by the
14461 assembler and the linker alone without help from the compiler.
14463 @item -mmcount-ra-address
14464 @itemx -mno-mcount-ra-address
14465 @opindex mmcount-ra-address
14466 @opindex mno-mcount-ra-address
14467 Emit (do not emit) code that allows @code{_mcount} to modify the
14468 calling function's return address. When enabled, this option extends
14469 the usual @code{_mcount} interface with a new @var{ra-address}
14470 parameter, which has type @code{intptr_t *} and is passed in register
14471 @code{$12}. @code{_mcount} can then modify the return address by
14472 doing both of the following:
14475 Returning the new address in register @code{$31}.
14477 Storing the new address in @code{*@var{ra-address}},
14478 if @var{ra-address} is nonnull.
14481 The default is @option{-mno-mcount-ra-address}.
14486 @subsection MMIX Options
14487 @cindex MMIX Options
14489 These options are defined for the MMIX:
14493 @itemx -mno-libfuncs
14495 @opindex mno-libfuncs
14496 Specify that intrinsic library functions are being compiled, passing all
14497 values in registers, no matter the size.
14500 @itemx -mno-epsilon
14502 @opindex mno-epsilon
14503 Generate floating-point comparison instructions that compare with respect
14504 to the @code{rE} epsilon register.
14506 @item -mabi=mmixware
14508 @opindex mabi=mmixware
14510 Generate code that passes function parameters and return values that (in
14511 the called function) are seen as registers @code{$0} and up, as opposed to
14512 the GNU ABI which uses global registers @code{$231} and up.
14514 @item -mzero-extend
14515 @itemx -mno-zero-extend
14516 @opindex mzero-extend
14517 @opindex mno-zero-extend
14518 When reading data from memory in sizes shorter than 64 bits, use (do not
14519 use) zero-extending load instructions by default, rather than
14520 sign-extending ones.
14523 @itemx -mno-knuthdiv
14525 @opindex mno-knuthdiv
14526 Make the result of a division yielding a remainder have the same sign as
14527 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
14528 remainder follows the sign of the dividend. Both methods are
14529 arithmetically valid, the latter being almost exclusively used.
14531 @item -mtoplevel-symbols
14532 @itemx -mno-toplevel-symbols
14533 @opindex mtoplevel-symbols
14534 @opindex mno-toplevel-symbols
14535 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
14536 code can be used with the @code{PREFIX} assembly directive.
14540 Generate an executable in the ELF format, rather than the default
14541 @samp{mmo} format used by the @command{mmix} simulator.
14543 @item -mbranch-predict
14544 @itemx -mno-branch-predict
14545 @opindex mbranch-predict
14546 @opindex mno-branch-predict
14547 Use (do not use) the probable-branch instructions, when static branch
14548 prediction indicates a probable branch.
14550 @item -mbase-addresses
14551 @itemx -mno-base-addresses
14552 @opindex mbase-addresses
14553 @opindex mno-base-addresses
14554 Generate (do not generate) code that uses @emph{base addresses}. Using a
14555 base address automatically generates a request (handled by the assembler
14556 and the linker) for a constant to be set up in a global register. The
14557 register is used for one or more base address requests within the range 0
14558 to 255 from the value held in the register. The generally leads to short
14559 and fast code, but the number of different data items that can be
14560 addressed is limited. This means that a program that uses lots of static
14561 data may require @option{-mno-base-addresses}.
14563 @item -msingle-exit
14564 @itemx -mno-single-exit
14565 @opindex msingle-exit
14566 @opindex mno-single-exit
14567 Force (do not force) generated code to have a single exit point in each
14571 @node MN10300 Options
14572 @subsection MN10300 Options
14573 @cindex MN10300 options
14575 These @option{-m} options are defined for Matsushita MN10300 architectures:
14580 Generate code to avoid bugs in the multiply instructions for the MN10300
14581 processors. This is the default.
14583 @item -mno-mult-bug
14584 @opindex mno-mult-bug
14585 Do not generate code to avoid bugs in the multiply instructions for the
14586 MN10300 processors.
14590 Generate code which uses features specific to the AM33 processor.
14594 Do not generate code which uses features specific to the AM33 processor. This
14597 @item -mreturn-pointer-on-d0
14598 @opindex mreturn-pointer-on-d0
14599 When generating a function which returns a pointer, return the pointer
14600 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
14601 only in a0, and attempts to call such functions without a prototype
14602 would result in errors. Note that this option is on by default; use
14603 @option{-mno-return-pointer-on-d0} to disable it.
14607 Do not link in the C run-time initialization object file.
14611 Indicate to the linker that it should perform a relaxation optimization pass
14612 to shorten branches, calls and absolute memory addresses. This option only
14613 has an effect when used on the command line for the final link step.
14615 This option makes symbolic debugging impossible.
14618 @node PDP-11 Options
14619 @subsection PDP-11 Options
14620 @cindex PDP-11 Options
14622 These options are defined for the PDP-11:
14627 Use hardware FPP floating point. This is the default. (FIS floating
14628 point on the PDP-11/40 is not supported.)
14631 @opindex msoft-float
14632 Do not use hardware floating point.
14636 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
14640 Return floating-point results in memory. This is the default.
14644 Generate code for a PDP-11/40.
14648 Generate code for a PDP-11/45. This is the default.
14652 Generate code for a PDP-11/10.
14654 @item -mbcopy-builtin
14655 @opindex mbcopy-builtin
14656 Use inline @code{movmemhi} patterns for copying memory. This is the
14661 Do not use inline @code{movmemhi} patterns for copying memory.
14667 Use 16-bit @code{int}. This is the default.
14673 Use 32-bit @code{int}.
14676 @itemx -mno-float32
14678 @opindex mno-float32
14679 Use 64-bit @code{float}. This is the default.
14682 @itemx -mno-float64
14684 @opindex mno-float64
14685 Use 32-bit @code{float}.
14689 Use @code{abshi2} pattern. This is the default.
14693 Do not use @code{abshi2} pattern.
14695 @item -mbranch-expensive
14696 @opindex mbranch-expensive
14697 Pretend that branches are expensive. This is for experimenting with
14698 code generation only.
14700 @item -mbranch-cheap
14701 @opindex mbranch-cheap
14702 Do not pretend that branches are expensive. This is the default.
14706 Generate code for a system with split I&D@.
14710 Generate code for a system without split I&D@. This is the default.
14714 Use Unix assembler syntax. This is the default when configured for
14715 @samp{pdp11-*-bsd}.
14719 Use DEC assembler syntax. This is the default when configured for any
14720 PDP-11 target other than @samp{pdp11-*-bsd}.
14723 @node picoChip Options
14724 @subsection picoChip Options
14725 @cindex picoChip options
14727 These @samp{-m} options are defined for picoChip implementations:
14731 @item -mae=@var{ae_type}
14733 Set the instruction set, register set, and instruction scheduling
14734 parameters for array element type @var{ae_type}. Supported values
14735 for @var{ae_type} are @samp{ANY}, @samp{MUL}, and @samp{MAC}.
14737 @option{-mae=ANY} selects a completely generic AE type. Code
14738 generated with this option will run on any of the other AE types. The
14739 code will not be as efficient as it would be if compiled for a specific
14740 AE type, and some types of operation (e.g., multiplication) will not
14741 work properly on all types of AE.
14743 @option{-mae=MUL} selects a MUL AE type. This is the most useful AE type
14744 for compiled code, and is the default.
14746 @option{-mae=MAC} selects a DSP-style MAC AE. Code compiled with this
14747 option may suffer from poor performance of byte (char) manipulation,
14748 since the DSP AE does not provide hardware support for byte load/stores.
14750 @item -msymbol-as-address
14751 Enable the compiler to directly use a symbol name as an address in a
14752 load/store instruction, without first loading it into a
14753 register. Typically, the use of this option will generate larger
14754 programs, which run faster than when the option isn't used. However, the
14755 results vary from program to program, so it is left as a user option,
14756 rather than being permanently enabled.
14758 @item -mno-inefficient-warnings
14759 Disables warnings about the generation of inefficient code. These
14760 warnings can be generated, for example, when compiling code which
14761 performs byte-level memory operations on the MAC AE type. The MAC AE has
14762 no hardware support for byte-level memory operations, so all byte
14763 load/stores must be synthesized from word load/store operations. This is
14764 inefficient and a warning will be generated indicating to the programmer
14765 that they should rewrite the code to avoid byte operations, or to target
14766 an AE type which has the necessary hardware support. This option enables
14767 the warning to be turned off.
14771 @node PowerPC Options
14772 @subsection PowerPC Options
14773 @cindex PowerPC options
14775 These are listed under @xref{RS/6000 and PowerPC Options}.
14777 @node RS/6000 and PowerPC Options
14778 @subsection IBM RS/6000 and PowerPC Options
14779 @cindex RS/6000 and PowerPC Options
14780 @cindex IBM RS/6000 and PowerPC Options
14782 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
14789 @itemx -mno-powerpc
14790 @itemx -mpowerpc-gpopt
14791 @itemx -mno-powerpc-gpopt
14792 @itemx -mpowerpc-gfxopt
14793 @itemx -mno-powerpc-gfxopt
14795 @itemx -mno-powerpc64
14799 @itemx -mno-popcntb
14801 @itemx -mno-popcntd
14809 @itemx -mno-hard-dfp
14813 @opindex mno-power2
14815 @opindex mno-powerpc
14816 @opindex mpowerpc-gpopt
14817 @opindex mno-powerpc-gpopt
14818 @opindex mpowerpc-gfxopt
14819 @opindex mno-powerpc-gfxopt
14820 @opindex mpowerpc64
14821 @opindex mno-powerpc64
14825 @opindex mno-popcntb
14827 @opindex mno-popcntd
14833 @opindex mno-mfpgpr
14835 @opindex mno-hard-dfp
14836 GCC supports two related instruction set architectures for the
14837 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
14838 instructions supported by the @samp{rios} chip set used in the original
14839 RS/6000 systems and the @dfn{PowerPC} instruction set is the
14840 architecture of the Freescale MPC5xx, MPC6xx, MPC8xx microprocessors, and
14841 the IBM 4xx, 6xx, and follow-on microprocessors.
14843 Neither architecture is a subset of the other. However there is a
14844 large common subset of instructions supported by both. An MQ
14845 register is included in processors supporting the POWER architecture.
14847 You use these options to specify which instructions are available on the
14848 processor you are using. The default value of these options is
14849 determined when configuring GCC@. Specifying the
14850 @option{-mcpu=@var{cpu_type}} overrides the specification of these
14851 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
14852 rather than the options listed above.
14854 The @option{-mpower} option allows GCC to generate instructions that
14855 are found only in the POWER architecture and to use the MQ register.
14856 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
14857 to generate instructions that are present in the POWER2 architecture but
14858 not the original POWER architecture.
14860 The @option{-mpowerpc} option allows GCC to generate instructions that
14861 are found only in the 32-bit subset of the PowerPC architecture.
14862 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
14863 GCC to use the optional PowerPC architecture instructions in the
14864 General Purpose group, including floating-point square root. Specifying
14865 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
14866 use the optional PowerPC architecture instructions in the Graphics
14867 group, including floating-point select.
14869 The @option{-mmfcrf} option allows GCC to generate the move from
14870 condition register field instruction implemented on the POWER4
14871 processor and other processors that support the PowerPC V2.01
14873 The @option{-mpopcntb} option allows GCC to generate the popcount and
14874 double precision FP reciprocal estimate instruction implemented on the
14875 POWER5 processor and other processors that support the PowerPC V2.02
14877 The @option{-mpopcntd} option allows GCC to generate the popcount
14878 instruction implemented on the POWER7 processor and other processors
14879 that support the PowerPC V2.06 architecture.
14880 The @option{-mfprnd} option allows GCC to generate the FP round to
14881 integer instructions implemented on the POWER5+ processor and other
14882 processors that support the PowerPC V2.03 architecture.
14883 The @option{-mcmpb} option allows GCC to generate the compare bytes
14884 instruction implemented on the POWER6 processor and other processors
14885 that support the PowerPC V2.05 architecture.
14886 The @option{-mmfpgpr} option allows GCC to generate the FP move to/from
14887 general purpose register instructions implemented on the POWER6X
14888 processor and other processors that support the extended PowerPC V2.05
14890 The @option{-mhard-dfp} option allows GCC to generate the decimal floating
14891 point instructions implemented on some POWER processors.
14893 The @option{-mpowerpc64} option allows GCC to generate the additional
14894 64-bit instructions that are found in the full PowerPC64 architecture
14895 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
14896 @option{-mno-powerpc64}.
14898 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
14899 will use only the instructions in the common subset of both
14900 architectures plus some special AIX common-mode calls, and will not use
14901 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
14902 permits GCC to use any instruction from either architecture and to
14903 allow use of the MQ register; specify this for the Motorola MPC601.
14905 @item -mnew-mnemonics
14906 @itemx -mold-mnemonics
14907 @opindex mnew-mnemonics
14908 @opindex mold-mnemonics
14909 Select which mnemonics to use in the generated assembler code. With
14910 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
14911 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
14912 assembler mnemonics defined for the POWER architecture. Instructions
14913 defined in only one architecture have only one mnemonic; GCC uses that
14914 mnemonic irrespective of which of these options is specified.
14916 GCC defaults to the mnemonics appropriate for the architecture in
14917 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
14918 value of these option. Unless you are building a cross-compiler, you
14919 should normally not specify either @option{-mnew-mnemonics} or
14920 @option{-mold-mnemonics}, but should instead accept the default.
14922 @item -mcpu=@var{cpu_type}
14924 Set architecture type, register usage, choice of mnemonics, and
14925 instruction scheduling parameters for machine type @var{cpu_type}.
14926 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
14927 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{464}, @samp{464fp},
14928 @samp{476}, @samp{476fp}, @samp{505}, @samp{601}, @samp{602}, @samp{603},
14929 @samp{603e}, @samp{604}, @samp{604e}, @samp{620}, @samp{630}, @samp{740},
14930 @samp{7400}, @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
14931 @samp{860}, @samp{970}, @samp{8540}, @samp{a2}, @samp{e300c2},
14932 @samp{e300c3}, @samp{e500mc}, @samp{e500mc64}, @samp{ec603e}, @samp{G3},
14933 @samp{G4}, @samp{G5}, @samp{power}, @samp{power2}, @samp{power3},
14934 @samp{power4}, @samp{power5}, @samp{power5+}, @samp{power6}, @samp{power6x},
14935 @samp{power7}, @samp{common}, @samp{powerpc}, @samp{powerpc64}, @samp{rios},
14936 @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64}.
14938 @option{-mcpu=common} selects a completely generic processor. Code
14939 generated under this option will run on any POWER or PowerPC processor.
14940 GCC will use only the instructions in the common subset of both
14941 architectures, and will not use the MQ register. GCC assumes a generic
14942 processor model for scheduling purposes.
14944 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
14945 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
14946 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
14947 types, with an appropriate, generic processor model assumed for
14948 scheduling purposes.
14950 The other options specify a specific processor. Code generated under
14951 those options will run best on that processor, and may not run at all on
14954 The @option{-mcpu} options automatically enable or disable the
14957 @gccoptlist{-maltivec -mfprnd -mhard-float -mmfcrf -mmultiple @gol
14958 -mnew-mnemonics -mpopcntb -mpopcntd -mpower -mpower2 -mpowerpc64 @gol
14959 -mpowerpc-gpopt -mpowerpc-gfxopt -msingle-float -mdouble-float @gol
14960 -msimple-fpu -mstring -mmulhw -mdlmzb -mmfpgpr -mvsx}
14962 The particular options set for any particular CPU will vary between
14963 compiler versions, depending on what setting seems to produce optimal
14964 code for that CPU; it doesn't necessarily reflect the actual hardware's
14965 capabilities. If you wish to set an individual option to a particular
14966 value, you may specify it after the @option{-mcpu} option, like
14967 @samp{-mcpu=970 -mno-altivec}.
14969 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
14970 not enabled or disabled by the @option{-mcpu} option at present because
14971 AIX does not have full support for these options. You may still
14972 enable or disable them individually if you're sure it'll work in your
14975 @item -mtune=@var{cpu_type}
14977 Set the instruction scheduling parameters for machine type
14978 @var{cpu_type}, but do not set the architecture type, register usage, or
14979 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
14980 values for @var{cpu_type} are used for @option{-mtune} as for
14981 @option{-mcpu}. If both are specified, the code generated will use the
14982 architecture, registers, and mnemonics set by @option{-mcpu}, but the
14983 scheduling parameters set by @option{-mtune}.
14985 @item -mcmodel=small
14986 @opindex mcmodel=small
14987 Generate PowerPC64 code for the small model: The TOC is limited to
14990 @item -mcmodel=large
14991 @opindex mcmodel=large
14992 Generate PowerPC64 code for the large model: The TOC may be up to 4G
14993 in size. Other data and code is only limited by the 64-bit address
14997 @itemx -mno-altivec
14999 @opindex mno-altivec
15000 Generate code that uses (does not use) AltiVec instructions, and also
15001 enable the use of built-in functions that allow more direct access to
15002 the AltiVec instruction set. You may also need to set
15003 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
15009 @opindex mno-vrsave
15010 Generate VRSAVE instructions when generating AltiVec code.
15012 @item -mgen-cell-microcode
15013 @opindex mgen-cell-microcode
15014 Generate Cell microcode instructions
15016 @item -mwarn-cell-microcode
15017 @opindex mwarn-cell-microcode
15018 Warning when a Cell microcode instruction is going to emitted. An example
15019 of a Cell microcode instruction is a variable shift.
15022 @opindex msecure-plt
15023 Generate code that allows ld and ld.so to build executables and shared
15024 libraries with non-exec .plt and .got sections. This is a PowerPC
15025 32-bit SYSV ABI option.
15029 Generate code that uses a BSS .plt section that ld.so fills in, and
15030 requires .plt and .got sections that are both writable and executable.
15031 This is a PowerPC 32-bit SYSV ABI option.
15037 This switch enables or disables the generation of ISEL instructions.
15039 @item -misel=@var{yes/no}
15040 This switch has been deprecated. Use @option{-misel} and
15041 @option{-mno-isel} instead.
15047 This switch enables or disables the generation of SPE simd
15053 @opindex mno-paired
15054 This switch enables or disables the generation of PAIRED simd
15057 @item -mspe=@var{yes/no}
15058 This option has been deprecated. Use @option{-mspe} and
15059 @option{-mno-spe} instead.
15065 Generate code that uses (does not use) vector/scalar (VSX)
15066 instructions, and also enable the use of built-in functions that allow
15067 more direct access to the VSX instruction set.
15069 @item -mfloat-gprs=@var{yes/single/double/no}
15070 @itemx -mfloat-gprs
15071 @opindex mfloat-gprs
15072 This switch enables or disables the generation of floating point
15073 operations on the general purpose registers for architectures that
15076 The argument @var{yes} or @var{single} enables the use of
15077 single-precision floating point operations.
15079 The argument @var{double} enables the use of single and
15080 double-precision floating point operations.
15082 The argument @var{no} disables floating point operations on the
15083 general purpose registers.
15085 This option is currently only available on the MPC854x.
15091 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
15092 targets (including GNU/Linux). The 32-bit environment sets int, long
15093 and pointer to 32 bits and generates code that runs on any PowerPC
15094 variant. The 64-bit environment sets int to 32 bits and long and
15095 pointer to 64 bits, and generates code for PowerPC64, as for
15096 @option{-mpowerpc64}.
15099 @itemx -mno-fp-in-toc
15100 @itemx -mno-sum-in-toc
15101 @itemx -mminimal-toc
15103 @opindex mno-fp-in-toc
15104 @opindex mno-sum-in-toc
15105 @opindex mminimal-toc
15106 Modify generation of the TOC (Table Of Contents), which is created for
15107 every executable file. The @option{-mfull-toc} option is selected by
15108 default. In that case, GCC will allocate at least one TOC entry for
15109 each unique non-automatic variable reference in your program. GCC
15110 will also place floating-point constants in the TOC@. However, only
15111 16,384 entries are available in the TOC@.
15113 If you receive a linker error message that saying you have overflowed
15114 the available TOC space, you can reduce the amount of TOC space used
15115 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
15116 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
15117 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
15118 generate code to calculate the sum of an address and a constant at
15119 run-time instead of putting that sum into the TOC@. You may specify one
15120 or both of these options. Each causes GCC to produce very slightly
15121 slower and larger code at the expense of conserving TOC space.
15123 If you still run out of space in the TOC even when you specify both of
15124 these options, specify @option{-mminimal-toc} instead. This option causes
15125 GCC to make only one TOC entry for every file. When you specify this
15126 option, GCC will produce code that is slower and larger but which
15127 uses extremely little TOC space. You may wish to use this option
15128 only on files that contain less frequently executed code.
15134 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
15135 @code{long} type, and the infrastructure needed to support them.
15136 Specifying @option{-maix64} implies @option{-mpowerpc64} and
15137 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
15138 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
15141 @itemx -mno-xl-compat
15142 @opindex mxl-compat
15143 @opindex mno-xl-compat
15144 Produce code that conforms more closely to IBM XL compiler semantics
15145 when using AIX-compatible ABI@. Pass floating-point arguments to
15146 prototyped functions beyond the register save area (RSA) on the stack
15147 in addition to argument FPRs. Do not assume that most significant
15148 double in 128-bit long double value is properly rounded when comparing
15149 values and converting to double. Use XL symbol names for long double
15152 The AIX calling convention was extended but not initially documented to
15153 handle an obscure K&R C case of calling a function that takes the
15154 address of its arguments with fewer arguments than declared. IBM XL
15155 compilers access floating point arguments which do not fit in the
15156 RSA from the stack when a subroutine is compiled without
15157 optimization. Because always storing floating-point arguments on the
15158 stack is inefficient and rarely needed, this option is not enabled by
15159 default and only is necessary when calling subroutines compiled by IBM
15160 XL compilers without optimization.
15164 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
15165 application written to use message passing with special startup code to
15166 enable the application to run. The system must have PE installed in the
15167 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
15168 must be overridden with the @option{-specs=} option to specify the
15169 appropriate directory location. The Parallel Environment does not
15170 support threads, so the @option{-mpe} option and the @option{-pthread}
15171 option are incompatible.
15173 @item -malign-natural
15174 @itemx -malign-power
15175 @opindex malign-natural
15176 @opindex malign-power
15177 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
15178 @option{-malign-natural} overrides the ABI-defined alignment of larger
15179 types, such as floating-point doubles, on their natural size-based boundary.
15180 The option @option{-malign-power} instructs GCC to follow the ABI-specified
15181 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
15183 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
15187 @itemx -mhard-float
15188 @opindex msoft-float
15189 @opindex mhard-float
15190 Generate code that does not use (uses) the floating-point register set.
15191 Software floating point emulation is provided if you use the
15192 @option{-msoft-float} option, and pass the option to GCC when linking.
15194 @item -msingle-float
15195 @itemx -mdouble-float
15196 @opindex msingle-float
15197 @opindex mdouble-float
15198 Generate code for single or double-precision floating point operations.
15199 @option{-mdouble-float} implies @option{-msingle-float}.
15202 @opindex msimple-fpu
15203 Do not generate sqrt and div instructions for hardware floating point unit.
15207 Specify type of floating point unit. Valid values are @var{sp_lite}
15208 (equivalent to -msingle-float -msimple-fpu), @var{dp_lite} (equivalent
15209 to -mdouble-float -msimple-fpu), @var{sp_full} (equivalent to -msingle-float),
15210 and @var{dp_full} (equivalent to -mdouble-float).
15213 @opindex mxilinx-fpu
15214 Perform optimizations for floating point unit on Xilinx PPC 405/440.
15217 @itemx -mno-multiple
15219 @opindex mno-multiple
15220 Generate code that uses (does not use) the load multiple word
15221 instructions and the store multiple word instructions. These
15222 instructions are generated by default on POWER systems, and not
15223 generated on PowerPC systems. Do not use @option{-mmultiple} on little
15224 endian PowerPC systems, since those instructions do not work when the
15225 processor is in little endian mode. The exceptions are PPC740 and
15226 PPC750 which permit the instructions usage in little endian mode.
15231 @opindex mno-string
15232 Generate code that uses (does not use) the load string instructions
15233 and the store string word instructions to save multiple registers and
15234 do small block moves. These instructions are generated by default on
15235 POWER systems, and not generated on PowerPC systems. Do not use
15236 @option{-mstring} on little endian PowerPC systems, since those
15237 instructions do not work when the processor is in little endian mode.
15238 The exceptions are PPC740 and PPC750 which permit the instructions
15239 usage in little endian mode.
15244 @opindex mno-update
15245 Generate code that uses (does not use) the load or store instructions
15246 that update the base register to the address of the calculated memory
15247 location. These instructions are generated by default. If you use
15248 @option{-mno-update}, there is a small window between the time that the
15249 stack pointer is updated and the address of the previous frame is
15250 stored, which means code that walks the stack frame across interrupts or
15251 signals may get corrupted data.
15253 @item -mavoid-indexed-addresses
15254 @itemx -mno-avoid-indexed-addresses
15255 @opindex mavoid-indexed-addresses
15256 @opindex mno-avoid-indexed-addresses
15257 Generate code that tries to avoid (not avoid) the use of indexed load
15258 or store instructions. These instructions can incur a performance
15259 penalty on Power6 processors in certain situations, such as when
15260 stepping through large arrays that cross a 16M boundary. This option
15261 is enabled by default when targetting Power6 and disabled otherwise.
15264 @itemx -mno-fused-madd
15265 @opindex mfused-madd
15266 @opindex mno-fused-madd
15267 Generate code that uses (does not use) the floating point multiply and
15268 accumulate instructions. These instructions are generated by default if
15269 hardware floating is used.
15275 Generate code that uses (does not use) the half-word multiply and
15276 multiply-accumulate instructions on the IBM 405, 440, 464 and 476 processors.
15277 These instructions are generated by default when targetting those
15284 Generate code that uses (does not use) the string-search @samp{dlmzb}
15285 instruction on the IBM 405, 440, 464 and 476 processors. This instruction is
15286 generated by default when targetting those processors.
15288 @item -mno-bit-align
15290 @opindex mno-bit-align
15291 @opindex mbit-align
15292 On System V.4 and embedded PowerPC systems do not (do) force structures
15293 and unions that contain bit-fields to be aligned to the base type of the
15296 For example, by default a structure containing nothing but 8
15297 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
15298 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
15299 the structure would be aligned to a 1 byte boundary and be one byte in
15302 @item -mno-strict-align
15303 @itemx -mstrict-align
15304 @opindex mno-strict-align
15305 @opindex mstrict-align
15306 On System V.4 and embedded PowerPC systems do not (do) assume that
15307 unaligned memory references will be handled by the system.
15309 @item -mrelocatable
15310 @itemx -mno-relocatable
15311 @opindex mrelocatable
15312 @opindex mno-relocatable
15313 On embedded PowerPC systems generate code that allows (does not allow)
15314 the program to be relocated to a different address at runtime. If you
15315 use @option{-mrelocatable} on any module, all objects linked together must
15316 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
15318 @item -mrelocatable-lib
15319 @itemx -mno-relocatable-lib
15320 @opindex mrelocatable-lib
15321 @opindex mno-relocatable-lib
15322 On embedded PowerPC systems generate code that allows (does not allow)
15323 the program to be relocated to a different address at runtime. Modules
15324 compiled with @option{-mrelocatable-lib} can be linked with either modules
15325 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
15326 with modules compiled with the @option{-mrelocatable} options.
15332 On System V.4 and embedded PowerPC systems do not (do) assume that
15333 register 2 contains a pointer to a global area pointing to the addresses
15334 used in the program.
15337 @itemx -mlittle-endian
15339 @opindex mlittle-endian
15340 On System V.4 and embedded PowerPC systems compile code for the
15341 processor in little endian mode. The @option{-mlittle-endian} option is
15342 the same as @option{-mlittle}.
15345 @itemx -mbig-endian
15347 @opindex mbig-endian
15348 On System V.4 and embedded PowerPC systems compile code for the
15349 processor in big endian mode. The @option{-mbig-endian} option is
15350 the same as @option{-mbig}.
15352 @item -mdynamic-no-pic
15353 @opindex mdynamic-no-pic
15354 On Darwin and Mac OS X systems, compile code so that it is not
15355 relocatable, but that its external references are relocatable. The
15356 resulting code is suitable for applications, but not shared
15359 @item -mprioritize-restricted-insns=@var{priority}
15360 @opindex mprioritize-restricted-insns
15361 This option controls the priority that is assigned to
15362 dispatch-slot restricted instructions during the second scheduling
15363 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
15364 @var{no/highest/second-highest} priority to dispatch slot restricted
15367 @item -msched-costly-dep=@var{dependence_type}
15368 @opindex msched-costly-dep
15369 This option controls which dependences are considered costly
15370 by the target during instruction scheduling. The argument
15371 @var{dependence_type} takes one of the following values:
15372 @var{no}: no dependence is costly,
15373 @var{all}: all dependences are costly,
15374 @var{true_store_to_load}: a true dependence from store to load is costly,
15375 @var{store_to_load}: any dependence from store to load is costly,
15376 @var{number}: any dependence which latency >= @var{number} is costly.
15378 @item -minsert-sched-nops=@var{scheme}
15379 @opindex minsert-sched-nops
15380 This option controls which nop insertion scheme will be used during
15381 the second scheduling pass. The argument @var{scheme} takes one of the
15383 @var{no}: Don't insert nops.
15384 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
15385 according to the scheduler's grouping.
15386 @var{regroup_exact}: Insert nops to force costly dependent insns into
15387 separate groups. Insert exactly as many nops as needed to force an insn
15388 to a new group, according to the estimated processor grouping.
15389 @var{number}: Insert nops to force costly dependent insns into
15390 separate groups. Insert @var{number} nops to force an insn to a new group.
15393 @opindex mcall-sysv
15394 On System V.4 and embedded PowerPC systems compile code using calling
15395 conventions that adheres to the March 1995 draft of the System V
15396 Application Binary Interface, PowerPC processor supplement. This is the
15397 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
15399 @item -mcall-sysv-eabi
15401 @opindex mcall-sysv-eabi
15402 @opindex mcall-eabi
15403 Specify both @option{-mcall-sysv} and @option{-meabi} options.
15405 @item -mcall-sysv-noeabi
15406 @opindex mcall-sysv-noeabi
15407 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
15409 @item -mcall-aixdesc
15411 On System V.4 and embedded PowerPC systems compile code for the AIX
15415 @opindex mcall-linux
15416 On System V.4 and embedded PowerPC systems compile code for the
15417 Linux-based GNU system.
15421 On System V.4 and embedded PowerPC systems compile code for the
15422 Hurd-based GNU system.
15424 @item -mcall-freebsd
15425 @opindex mcall-freebsd
15426 On System V.4 and embedded PowerPC systems compile code for the
15427 FreeBSD operating system.
15429 @item -mcall-netbsd
15430 @opindex mcall-netbsd
15431 On System V.4 and embedded PowerPC systems compile code for the
15432 NetBSD operating system.
15434 @item -mcall-openbsd
15435 @opindex mcall-netbsd
15436 On System V.4 and embedded PowerPC systems compile code for the
15437 OpenBSD operating system.
15439 @item -maix-struct-return
15440 @opindex maix-struct-return
15441 Return all structures in memory (as specified by the AIX ABI)@.
15443 @item -msvr4-struct-return
15444 @opindex msvr4-struct-return
15445 Return structures smaller than 8 bytes in registers (as specified by the
15448 @item -mabi=@var{abi-type}
15450 Extend the current ABI with a particular extension, or remove such extension.
15451 Valid values are @var{altivec}, @var{no-altivec}, @var{spe},
15452 @var{no-spe}, @var{ibmlongdouble}, @var{ieeelongdouble}@.
15456 Extend the current ABI with SPE ABI extensions. This does not change
15457 the default ABI, instead it adds the SPE ABI extensions to the current
15461 @opindex mabi=no-spe
15462 Disable Booke SPE ABI extensions for the current ABI@.
15464 @item -mabi=ibmlongdouble
15465 @opindex mabi=ibmlongdouble
15466 Change the current ABI to use IBM extended precision long double.
15467 This is a PowerPC 32-bit SYSV ABI option.
15469 @item -mabi=ieeelongdouble
15470 @opindex mabi=ieeelongdouble
15471 Change the current ABI to use IEEE extended precision long double.
15472 This is a PowerPC 32-bit Linux ABI option.
15475 @itemx -mno-prototype
15476 @opindex mprototype
15477 @opindex mno-prototype
15478 On System V.4 and embedded PowerPC systems assume that all calls to
15479 variable argument functions are properly prototyped. Otherwise, the
15480 compiler must insert an instruction before every non prototyped call to
15481 set or clear bit 6 of the condition code register (@var{CR}) to
15482 indicate whether floating point values were passed in the floating point
15483 registers in case the function takes a variable arguments. With
15484 @option{-mprototype}, only calls to prototyped variable argument functions
15485 will set or clear the bit.
15489 On embedded PowerPC systems, assume that the startup module is called
15490 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
15491 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}
15496 On embedded PowerPC systems, assume that the startup module is called
15497 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
15502 On embedded PowerPC systems, assume that the startup module is called
15503 @file{crt0.o} and the standard C libraries are @file{libads.a} and
15506 @item -myellowknife
15507 @opindex myellowknife
15508 On embedded PowerPC systems, assume that the startup module is called
15509 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
15514 On System V.4 and embedded PowerPC systems, specify that you are
15515 compiling for a VxWorks system.
15519 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
15520 header to indicate that @samp{eabi} extended relocations are used.
15526 On System V.4 and embedded PowerPC systems do (do not) adhere to the
15527 Embedded Applications Binary Interface (eabi) which is a set of
15528 modifications to the System V.4 specifications. Selecting @option{-meabi}
15529 means that the stack is aligned to an 8 byte boundary, a function
15530 @code{__eabi} is called to from @code{main} to set up the eabi
15531 environment, and the @option{-msdata} option can use both @code{r2} and
15532 @code{r13} to point to two separate small data areas. Selecting
15533 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
15534 do not call an initialization function from @code{main}, and the
15535 @option{-msdata} option will only use @code{r13} to point to a single
15536 small data area. The @option{-meabi} option is on by default if you
15537 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
15540 @opindex msdata=eabi
15541 On System V.4 and embedded PowerPC systems, put small initialized
15542 @code{const} global and static data in the @samp{.sdata2} section, which
15543 is pointed to by register @code{r2}. Put small initialized
15544 non-@code{const} global and static data in the @samp{.sdata} section,
15545 which is pointed to by register @code{r13}. Put small uninitialized
15546 global and static data in the @samp{.sbss} section, which is adjacent to
15547 the @samp{.sdata} section. The @option{-msdata=eabi} option is
15548 incompatible with the @option{-mrelocatable} option. The
15549 @option{-msdata=eabi} option also sets the @option{-memb} option.
15552 @opindex msdata=sysv
15553 On System V.4 and embedded PowerPC systems, put small global and static
15554 data in the @samp{.sdata} section, which is pointed to by register
15555 @code{r13}. Put small uninitialized global and static data in the
15556 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
15557 The @option{-msdata=sysv} option is incompatible with the
15558 @option{-mrelocatable} option.
15560 @item -msdata=default
15562 @opindex msdata=default
15564 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
15565 compile code the same as @option{-msdata=eabi}, otherwise compile code the
15566 same as @option{-msdata=sysv}.
15569 @opindex msdata=data
15570 On System V.4 and embedded PowerPC systems, put small global
15571 data in the @samp{.sdata} section. Put small uninitialized global
15572 data in the @samp{.sbss} section. Do not use register @code{r13}
15573 to address small data however. This is the default behavior unless
15574 other @option{-msdata} options are used.
15578 @opindex msdata=none
15580 On embedded PowerPC systems, put all initialized global and static data
15581 in the @samp{.data} section, and all uninitialized data in the
15582 @samp{.bss} section.
15586 @cindex smaller data references (PowerPC)
15587 @cindex .sdata/.sdata2 references (PowerPC)
15588 On embedded PowerPC systems, put global and static items less than or
15589 equal to @var{num} bytes into the small data or bss sections instead of
15590 the normal data or bss section. By default, @var{num} is 8. The
15591 @option{-G @var{num}} switch is also passed to the linker.
15592 All modules should be compiled with the same @option{-G @var{num}} value.
15595 @itemx -mno-regnames
15597 @opindex mno-regnames
15598 On System V.4 and embedded PowerPC systems do (do not) emit register
15599 names in the assembly language output using symbolic forms.
15602 @itemx -mno-longcall
15604 @opindex mno-longcall
15605 By default assume that all calls are far away so that a longer more
15606 expensive calling sequence is required. This is required for calls
15607 further than 32 megabytes (33,554,432 bytes) from the current location.
15608 A short call will be generated if the compiler knows
15609 the call cannot be that far away. This setting can be overridden by
15610 the @code{shortcall} function attribute, or by @code{#pragma
15613 Some linkers are capable of detecting out-of-range calls and generating
15614 glue code on the fly. On these systems, long calls are unnecessary and
15615 generate slower code. As of this writing, the AIX linker can do this,
15616 as can the GNU linker for PowerPC/64. It is planned to add this feature
15617 to the GNU linker for 32-bit PowerPC systems as well.
15619 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
15620 callee, L42'', plus a ``branch island'' (glue code). The two target
15621 addresses represent the callee and the ``branch island''. The
15622 Darwin/PPC linker will prefer the first address and generate a ``bl
15623 callee'' if the PPC ``bl'' instruction will reach the callee directly;
15624 otherwise, the linker will generate ``bl L42'' to call the ``branch
15625 island''. The ``branch island'' is appended to the body of the
15626 calling function; it computes the full 32-bit address of the callee
15629 On Mach-O (Darwin) systems, this option directs the compiler emit to
15630 the glue for every direct call, and the Darwin linker decides whether
15631 to use or discard it.
15633 In the future, we may cause GCC to ignore all longcall specifications
15634 when the linker is known to generate glue.
15636 @item -mtls-markers
15637 @itemx -mno-tls-markers
15638 @opindex mtls-markers
15639 @opindex mno-tls-markers
15640 Mark (do not mark) calls to @code{__tls_get_addr} with a relocation
15641 specifying the function argument. The relocation allows ld to
15642 reliably associate function call with argument setup instructions for
15643 TLS optimization, which in turn allows gcc to better schedule the
15648 Adds support for multithreading with the @dfn{pthreads} library.
15649 This option sets flags for both the preprocessor and linker.
15654 This option will enable GCC to use the reciprocal estimate and
15655 reciprocal square root estimate instructions with additional
15656 Newton-Raphson steps to increase precision instead of doing a divide or
15657 square root and divide for floating point arguments. You should use
15658 the @option{-ffast-math} option when using @option{-mrecip} (or at
15659 least @option{-funsafe-math-optimizations},
15660 @option{-finite-math-only}, @option{-freciprocal-math} and
15661 @option{-fno-trapping-math}). Note that while the throughput of the
15662 sequence is generally higher than the throughput of the non-reciprocal
15663 instruction, the precision of the sequence can be decreased by up to 2
15664 ulp (i.e. the inverse of 1.0 equals 0.99999994) for reciprocal square
15667 @item -mrecip=@var{opt}
15668 @opindex mrecip=opt
15669 This option allows to control which reciprocal estimate instructions
15670 may be used. @var{opt} is a comma separated list of options, that may
15671 be preceeded by a @code{!} to invert the option:
15672 @code{all}: enable all estimate instructions,
15673 @code{default}: enable the default instructions, equvalent to @option{-mrecip},
15674 @code{none}: disable all estimate instructions, equivalent to @option{-mno-recip};
15675 @code{div}: enable the reciprocal approximation instructions for both single and double precision;
15676 @code{divf}: enable the single precision reciprocal approximation instructions;
15677 @code{divd}: enable the double precision reciprocal approximation instructions;
15678 @code{rsqrt}: enable the reciprocal square root approximation instructions for both single and double precision;
15679 @code{rsqrtf}: enable the single precision reciprocal square root approximation instructions;
15680 @code{rsqrtd}: enable the double precision reciprocal square root approximation instructions;
15682 So for example, @option{-mrecip=all,!rsqrtd} would enable the
15683 all of the reciprocal estimate instructions, except for the
15684 @code{FRSQRTE}, @code{XSRSQRTEDP}, and @code{XVRSQRTEDP} instructions
15685 which handle the double precision reciprocal square root calculations.
15687 @item -mrecip-precision
15688 @itemx -mno-recip-precision
15689 @opindex mrecip-precision
15690 Assume (do not assume) that the reciprocal estimate instructions
15691 provide higher precision estimates than is mandated by the powerpc
15692 ABI. Selecting @option{-mcpu=power6} or @option{-mcpu=power7}
15693 automatically selects @option{-mrecip-precision}. The double
15694 precision square root estimate instructions are not generated by
15695 default on low precision machines, since they do not provide an
15696 estimate that converges after three steps.
15700 @subsection RX Options
15703 These command line options are defined for RX targets:
15706 @item -m64bit-doubles
15707 @itemx -m32bit-doubles
15708 @opindex m64bit-doubles
15709 @opindex m32bit-doubles
15710 Make the @code{double} data type be 64-bits (@option{-m64bit-doubles})
15711 or 32-bits (@option{-m32bit-doubles}) in size. The default is
15712 @option{-m32bit-doubles}. @emph{Note} RX floating point hardware only
15713 works on 32-bit values, which is why the default is
15714 @option{-m32bit-doubles}.
15720 Enables (@option{-fpu}) or disables (@option{-nofpu}) the use of RX
15721 floating point hardware. The default is enabled for the @var{RX600}
15722 series and disabled for the @var{RX200} series.
15724 Floating point instructions will only be generated for 32-bit floating
15725 point values however, so if the @option{-m64bit-doubles} option is in
15726 use then the FPU hardware will not be used for doubles.
15728 @emph{Note} If the @option{-fpu} option is enabled then
15729 @option{-funsafe-math-optimizations} is also enabled automatically.
15730 This is because the RX FPU instructions are themselves unsafe.
15732 @item -mcpu=@var{name}
15733 @itemx -patch=@var{name}
15736 Selects the type of RX CPU to be targeted. Currently three types are
15737 supported, the generic @var{RX600} and @var{RX200} series hardware and
15738 the specific @var{RX610} cpu. The default is @var{RX600}.
15740 The only difference between @var{RX600} and @var{RX610} is that the
15741 @var{RX610} does not support the @code{MVTIPL} instruction.
15743 The @var{RX200} series does not have a hardware floating point unit
15744 and so @option{-nofpu} is enabled by default when this type is
15747 @item -mbig-endian-data
15748 @itemx -mlittle-endian-data
15749 @opindex mbig-endian-data
15750 @opindex mlittle-endian-data
15751 Store data (but not code) in the big-endian format. The default is
15752 @option{-mlittle-endian-data}, ie to store data in the little endian
15755 @item -msmall-data-limit=@var{N}
15756 @opindex msmall-data-limit
15757 Specifies the maximum size in bytes of global and static variables
15758 which can be placed into the small data area. Using the small data
15759 area can lead to smaller and faster code, but the size of area is
15760 limited and it is up to the programmer to ensure that the area does
15761 not overflow. Also when the small data area is used one of the RX's
15762 registers (@code{r13}) is reserved for use pointing to this area, so
15763 it is no longer available for use by the compiler. This could result
15764 in slower and/or larger code if variables which once could have been
15765 held in @code{r13} are now pushed onto the stack.
15767 Note, common variables (variables which have not been initialised) and
15768 constants are not placed into the small data area as they are assigned
15769 to other sections in the output executable.
15771 The default value is zero, which disables this feature. Note, this
15772 feature is not enabled by default with higher optimization levels
15773 (@option{-O2} etc) because of the potentially detrimental effects of
15774 reserving register @code{r13}. It is up to the programmer to
15775 experiment and discover whether this feature is of benefit to their
15782 Use the simulator runtime. The default is to use the libgloss board
15785 @item -mas100-syntax
15786 @itemx -mno-as100-syntax
15787 @opindex mas100-syntax
15788 @opindex mno-as100-syntax
15789 When generating assembler output use a syntax that is compatible with
15790 Renesas's AS100 assembler. This syntax can also be handled by the GAS
15791 assembler but it has some restrictions so generating it is not the
15794 @item -mmax-constant-size=@var{N}
15795 @opindex mmax-constant-size
15796 Specifies the maximum size, in bytes, of a constant that can be used as
15797 an operand in a RX instruction. Although the RX instruction set does
15798 allow constants of up to 4 bytes in length to be used in instructions,
15799 a longer value equates to a longer instruction. Thus in some
15800 circumstances it can be beneficial to restrict the size of constants
15801 that are used in instructions. Constants that are too big are instead
15802 placed into a constant pool and referenced via register indirection.
15804 The value @var{N} can be between 0 and 4. A value of 0 (the default)
15805 or 4 means that constants of any size are allowed.
15809 Enable linker relaxation. Linker relaxation is a process whereby the
15810 linker will attempt to reduce the size of a program by finding shorter
15811 versions of various instructions. Disabled by default.
15813 @item -mint-register=@var{N}
15814 @opindex mint-register
15815 Specify the number of registers to reserve for fast interrupt handler
15816 functions. The value @var{N} can be between 0 and 4. A value of 1
15817 means that register @code{r13} will be reserved for the exclusive use
15818 of fast interrupt handlers. A value of 2 reserves @code{r13} and
15819 @code{r12}. A value of 3 reserves @code{r13}, @code{r12} and
15820 @code{r11}, and a value of 4 reserves @code{r13} through @code{r10}.
15821 A value of 0, the default, does not reserve any registers.
15823 @item -msave-acc-in-interrupts
15824 @opindex msave-acc-in-interrupts
15825 Specifies that interrupt handler functions should preserve the
15826 accumulator register. This is only necessary if normal code might use
15827 the accumulator register, for example because it performs 64-bit
15828 multiplications. The default is to ignore the accumulator as this
15829 makes the interrupt handlers faster.
15833 @emph{Note:} The generic GCC command line @option{-ffixed-@var{reg}}
15834 has special significance to the RX port when used with the
15835 @code{interrupt} function attribute. This attribute indicates a
15836 function intended to process fast interrupts. GCC will will ensure
15837 that it only uses the registers @code{r10}, @code{r11}, @code{r12}
15838 and/or @code{r13} and only provided that the normal use of the
15839 corresponding registers have been restricted via the
15840 @option{-ffixed-@var{reg}} or @option{-mint-register} command line
15843 @node S/390 and zSeries Options
15844 @subsection S/390 and zSeries Options
15845 @cindex S/390 and zSeries Options
15847 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
15851 @itemx -msoft-float
15852 @opindex mhard-float
15853 @opindex msoft-float
15854 Use (do not use) the hardware floating-point instructions and registers
15855 for floating-point operations. When @option{-msoft-float} is specified,
15856 functions in @file{libgcc.a} will be used to perform floating-point
15857 operations. When @option{-mhard-float} is specified, the compiler
15858 generates IEEE floating-point instructions. This is the default.
15861 @itemx -mno-hard-dfp
15863 @opindex mno-hard-dfp
15864 Use (do not use) the hardware decimal-floating-point instructions for
15865 decimal-floating-point operations. When @option{-mno-hard-dfp} is
15866 specified, functions in @file{libgcc.a} will be used to perform
15867 decimal-floating-point operations. When @option{-mhard-dfp} is
15868 specified, the compiler generates decimal-floating-point hardware
15869 instructions. This is the default for @option{-march=z9-ec} or higher.
15871 @item -mlong-double-64
15872 @itemx -mlong-double-128
15873 @opindex mlong-double-64
15874 @opindex mlong-double-128
15875 These switches control the size of @code{long double} type. A size
15876 of 64bit makes the @code{long double} type equivalent to the @code{double}
15877 type. This is the default.
15880 @itemx -mno-backchain
15881 @opindex mbackchain
15882 @opindex mno-backchain
15883 Store (do not store) the address of the caller's frame as backchain pointer
15884 into the callee's stack frame.
15885 A backchain may be needed to allow debugging using tools that do not understand
15886 DWARF-2 call frame information.
15887 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
15888 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
15889 the backchain is placed into the topmost word of the 96/160 byte register
15892 In general, code compiled with @option{-mbackchain} is call-compatible with
15893 code compiled with @option{-mmo-backchain}; however, use of the backchain
15894 for debugging purposes usually requires that the whole binary is built with
15895 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
15896 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
15897 to build a linux kernel use @option{-msoft-float}.
15899 The default is to not maintain the backchain.
15901 @item -mpacked-stack
15902 @itemx -mno-packed-stack
15903 @opindex mpacked-stack
15904 @opindex mno-packed-stack
15905 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
15906 specified, the compiler uses the all fields of the 96/160 byte register save
15907 area only for their default purpose; unused fields still take up stack space.
15908 When @option{-mpacked-stack} is specified, register save slots are densely
15909 packed at the top of the register save area; unused space is reused for other
15910 purposes, allowing for more efficient use of the available stack space.
15911 However, when @option{-mbackchain} is also in effect, the topmost word of
15912 the save area is always used to store the backchain, and the return address
15913 register is always saved two words below the backchain.
15915 As long as the stack frame backchain is not used, code generated with
15916 @option{-mpacked-stack} is call-compatible with code generated with
15917 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
15918 S/390 or zSeries generated code that uses the stack frame backchain at run
15919 time, not just for debugging purposes. Such code is not call-compatible
15920 with code compiled with @option{-mpacked-stack}. Also, note that the
15921 combination of @option{-mbackchain},
15922 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
15923 to build a linux kernel use @option{-msoft-float}.
15925 The default is to not use the packed stack layout.
15928 @itemx -mno-small-exec
15929 @opindex msmall-exec
15930 @opindex mno-small-exec
15931 Generate (or do not generate) code using the @code{bras} instruction
15932 to do subroutine calls.
15933 This only works reliably if the total executable size does not
15934 exceed 64k. The default is to use the @code{basr} instruction instead,
15935 which does not have this limitation.
15941 When @option{-m31} is specified, generate code compliant to the
15942 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
15943 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
15944 particular to generate 64-bit instructions. For the @samp{s390}
15945 targets, the default is @option{-m31}, while the @samp{s390x}
15946 targets default to @option{-m64}.
15952 When @option{-mzarch} is specified, generate code using the
15953 instructions available on z/Architecture.
15954 When @option{-mesa} is specified, generate code using the
15955 instructions available on ESA/390. Note that @option{-mesa} is
15956 not possible with @option{-m64}.
15957 When generating code compliant to the GNU/Linux for S/390 ABI,
15958 the default is @option{-mesa}. When generating code compliant
15959 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
15965 Generate (or do not generate) code using the @code{mvcle} instruction
15966 to perform block moves. When @option{-mno-mvcle} is specified,
15967 use a @code{mvc} loop instead. This is the default unless optimizing for
15974 Print (or do not print) additional debug information when compiling.
15975 The default is to not print debug information.
15977 @item -march=@var{cpu-type}
15979 Generate code that will run on @var{cpu-type}, which is the name of a system
15980 representing a certain processor type. Possible values for
15981 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, @samp{z990},
15982 @samp{z9-109}, @samp{z9-ec} and @samp{z10}.
15983 When generating code using the instructions available on z/Architecture,
15984 the default is @option{-march=z900}. Otherwise, the default is
15985 @option{-march=g5}.
15987 @item -mtune=@var{cpu-type}
15989 Tune to @var{cpu-type} everything applicable about the generated code,
15990 except for the ABI and the set of available instructions.
15991 The list of @var{cpu-type} values is the same as for @option{-march}.
15992 The default is the value used for @option{-march}.
15995 @itemx -mno-tpf-trace
15996 @opindex mtpf-trace
15997 @opindex mno-tpf-trace
15998 Generate code that adds (does not add) in TPF OS specific branches to trace
15999 routines in the operating system. This option is off by default, even
16000 when compiling for the TPF OS@.
16003 @itemx -mno-fused-madd
16004 @opindex mfused-madd
16005 @opindex mno-fused-madd
16006 Generate code that uses (does not use) the floating point multiply and
16007 accumulate instructions. These instructions are generated by default if
16008 hardware floating point is used.
16010 @item -mwarn-framesize=@var{framesize}
16011 @opindex mwarn-framesize
16012 Emit a warning if the current function exceeds the given frame size. Because
16013 this is a compile time check it doesn't need to be a real problem when the program
16014 runs. It is intended to identify functions which most probably cause
16015 a stack overflow. It is useful to be used in an environment with limited stack
16016 size e.g.@: the linux kernel.
16018 @item -mwarn-dynamicstack
16019 @opindex mwarn-dynamicstack
16020 Emit a warning if the function calls alloca or uses dynamically
16021 sized arrays. This is generally a bad idea with a limited stack size.
16023 @item -mstack-guard=@var{stack-guard}
16024 @itemx -mstack-size=@var{stack-size}
16025 @opindex mstack-guard
16026 @opindex mstack-size
16027 If these options are provided the s390 back end emits additional instructions in
16028 the function prologue which trigger a trap if the stack size is @var{stack-guard}
16029 bytes above the @var{stack-size} (remember that the stack on s390 grows downward).
16030 If the @var{stack-guard} option is omitted the smallest power of 2 larger than
16031 the frame size of the compiled function is chosen.
16032 These options are intended to be used to help debugging stack overflow problems.
16033 The additionally emitted code causes only little overhead and hence can also be
16034 used in production like systems without greater performance degradation. The given
16035 values have to be exact powers of 2 and @var{stack-size} has to be greater than
16036 @var{stack-guard} without exceeding 64k.
16037 In order to be efficient the extra code makes the assumption that the stack starts
16038 at an address aligned to the value given by @var{stack-size}.
16039 The @var{stack-guard} option can only be used in conjunction with @var{stack-size}.
16042 @node Score Options
16043 @subsection Score Options
16044 @cindex Score Options
16046 These options are defined for Score implementations:
16051 Compile code for big endian mode. This is the default.
16055 Compile code for little endian mode.
16059 Disable generate bcnz instruction.
16063 Enable generate unaligned load and store instruction.
16067 Enable the use of multiply-accumulate instructions. Disabled by default.
16071 Specify the SCORE5 as the target architecture.
16075 Specify the SCORE5U of the target architecture.
16079 Specify the SCORE7 as the target architecture. This is the default.
16083 Specify the SCORE7D as the target architecture.
16087 @subsection SH Options
16089 These @samp{-m} options are defined for the SH implementations:
16094 Generate code for the SH1.
16098 Generate code for the SH2.
16101 Generate code for the SH2e.
16105 Generate code for the SH2a without FPU, or for a SH2a-FPU in such a way
16106 that the floating-point unit is not used.
16108 @item -m2a-single-only
16109 @opindex m2a-single-only
16110 Generate code for the SH2a-FPU, in such a way that no double-precision
16111 floating point operations are used.
16114 @opindex m2a-single
16115 Generate code for the SH2a-FPU assuming the floating-point unit is in
16116 single-precision mode by default.
16120 Generate code for the SH2a-FPU assuming the floating-point unit is in
16121 double-precision mode by default.
16125 Generate code for the SH3.
16129 Generate code for the SH3e.
16133 Generate code for the SH4 without a floating-point unit.
16135 @item -m4-single-only
16136 @opindex m4-single-only
16137 Generate code for the SH4 with a floating-point unit that only
16138 supports single-precision arithmetic.
16142 Generate code for the SH4 assuming the floating-point unit is in
16143 single-precision mode by default.
16147 Generate code for the SH4.
16151 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
16152 floating-point unit is not used.
16154 @item -m4a-single-only
16155 @opindex m4a-single-only
16156 Generate code for the SH4a, in such a way that no double-precision
16157 floating point operations are used.
16160 @opindex m4a-single
16161 Generate code for the SH4a assuming the floating-point unit is in
16162 single-precision mode by default.
16166 Generate code for the SH4a.
16170 Same as @option{-m4a-nofpu}, except that it implicitly passes
16171 @option{-dsp} to the assembler. GCC doesn't generate any DSP
16172 instructions at the moment.
16176 Compile code for the processor in big endian mode.
16180 Compile code for the processor in little endian mode.
16184 Align doubles at 64-bit boundaries. Note that this changes the calling
16185 conventions, and thus some functions from the standard C library will
16186 not work unless you recompile it first with @option{-mdalign}.
16190 Shorten some address references at link time, when possible; uses the
16191 linker option @option{-relax}.
16195 Use 32-bit offsets in @code{switch} tables. The default is to use
16200 Enable the use of bit manipulation instructions on SH2A.
16204 Enable the use of the instruction @code{fmovd}. Check @option{-mdalign} for
16205 alignment constraints.
16209 Comply with the calling conventions defined by Renesas.
16213 Comply with the calling conventions defined by Renesas.
16217 Comply with the calling conventions defined for GCC before the Renesas
16218 conventions were available. This option is the default for all
16219 targets of the SH toolchain except for @samp{sh-symbianelf}.
16222 @opindex mnomacsave
16223 Mark the @code{MAC} register as call-clobbered, even if
16224 @option{-mhitachi} is given.
16228 Increase IEEE-compliance of floating-point code.
16229 At the moment, this is equivalent to @option{-fno-finite-math-only}.
16230 When generating 16 bit SH opcodes, getting IEEE-conforming results for
16231 comparisons of NANs / infinities incurs extra overhead in every
16232 floating point comparison, therefore the default is set to
16233 @option{-ffinite-math-only}.
16235 @item -minline-ic_invalidate
16236 @opindex minline-ic_invalidate
16237 Inline code to invalidate instruction cache entries after setting up
16238 nested function trampolines.
16239 This option has no effect if -musermode is in effect and the selected
16240 code generation option (e.g. -m4) does not allow the use of the icbi
16242 If the selected code generation option does not allow the use of the icbi
16243 instruction, and -musermode is not in effect, the inlined code will
16244 manipulate the instruction cache address array directly with an associative
16245 write. This not only requires privileged mode, but it will also
16246 fail if the cache line had been mapped via the TLB and has become unmapped.
16250 Dump instruction size and location in the assembly code.
16253 @opindex mpadstruct
16254 This option is deprecated. It pads structures to multiple of 4 bytes,
16255 which is incompatible with the SH ABI@.
16259 Optimize for space instead of speed. Implied by @option{-Os}.
16262 @opindex mprefergot
16263 When generating position-independent code, emit function calls using
16264 the Global Offset Table instead of the Procedure Linkage Table.
16268 Don't generate privileged mode only code; implies -mno-inline-ic_invalidate
16269 if the inlined code would not work in user mode.
16270 This is the default when the target is @code{sh-*-linux*}.
16272 @item -multcost=@var{number}
16273 @opindex multcost=@var{number}
16274 Set the cost to assume for a multiply insn.
16276 @item -mdiv=@var{strategy}
16277 @opindex mdiv=@var{strategy}
16278 Set the division strategy to use for SHmedia code. @var{strategy} must be
16279 one of: call, call2, fp, inv, inv:minlat, inv20u, inv20l, inv:call,
16280 inv:call2, inv:fp .
16281 "fp" performs the operation in floating point. This has a very high latency,
16282 but needs only a few instructions, so it might be a good choice if
16283 your code has enough easily exploitable ILP to allow the compiler to
16284 schedule the floating point instructions together with other instructions.
16285 Division by zero causes a floating point exception.
16286 "inv" uses integer operations to calculate the inverse of the divisor,
16287 and then multiplies the dividend with the inverse. This strategy allows
16288 cse and hoisting of the inverse calculation. Division by zero calculates
16289 an unspecified result, but does not trap.
16290 "inv:minlat" is a variant of "inv" where if no cse / hoisting opportunities
16291 have been found, or if the entire operation has been hoisted to the same
16292 place, the last stages of the inverse calculation are intertwined with the
16293 final multiply to reduce the overall latency, at the expense of using a few
16294 more instructions, and thus offering fewer scheduling opportunities with
16296 "call" calls a library function that usually implements the inv:minlat
16298 This gives high code density for m5-*media-nofpu compilations.
16299 "call2" uses a different entry point of the same library function, where it
16300 assumes that a pointer to a lookup table has already been set up, which
16301 exposes the pointer load to cse / code hoisting optimizations.
16302 "inv:call", "inv:call2" and "inv:fp" all use the "inv" algorithm for initial
16303 code generation, but if the code stays unoptimized, revert to the "call",
16304 "call2", or "fp" strategies, respectively. Note that the
16305 potentially-trapping side effect of division by zero is carried by a
16306 separate instruction, so it is possible that all the integer instructions
16307 are hoisted out, but the marker for the side effect stays where it is.
16308 A recombination to fp operations or a call is not possible in that case.
16309 "inv20u" and "inv20l" are variants of the "inv:minlat" strategy. In the case
16310 that the inverse calculation was nor separated from the multiply, they speed
16311 up division where the dividend fits into 20 bits (plus sign where applicable),
16312 by inserting a test to skip a number of operations in this case; this test
16313 slows down the case of larger dividends. inv20u assumes the case of a such
16314 a small dividend to be unlikely, and inv20l assumes it to be likely.
16316 @item -maccumulate-outgoing-args
16317 @opindex maccumulate-outgoing-args
16318 Reserve space once for outgoing arguments in the function prologue rather
16319 than around each call. Generally beneficial for performance and size. Also
16320 needed for unwinding to avoid changing the stack frame around conditional code.
16322 @item -mdivsi3_libfunc=@var{name}
16323 @opindex mdivsi3_libfunc=@var{name}
16324 Set the name of the library function used for 32 bit signed division to
16325 @var{name}. This only affect the name used in the call and inv:call
16326 division strategies, and the compiler will still expect the same
16327 sets of input/output/clobbered registers as if this option was not present.
16329 @item -mfixed-range=@var{register-range}
16330 @opindex mfixed-range
16331 Generate code treating the given register range as fixed registers.
16332 A fixed register is one that the register allocator can not use. This is
16333 useful when compiling kernel code. A register range is specified as
16334 two registers separated by a dash. Multiple register ranges can be
16335 specified separated by a comma.
16337 @item -madjust-unroll
16338 @opindex madjust-unroll
16339 Throttle unrolling to avoid thrashing target registers.
16340 This option only has an effect if the gcc code base supports the
16341 TARGET_ADJUST_UNROLL_MAX target hook.
16343 @item -mindexed-addressing
16344 @opindex mindexed-addressing
16345 Enable the use of the indexed addressing mode for SHmedia32/SHcompact.
16346 This is only safe if the hardware and/or OS implement 32 bit wrap-around
16347 semantics for the indexed addressing mode. The architecture allows the
16348 implementation of processors with 64 bit MMU, which the OS could use to
16349 get 32 bit addressing, but since no current hardware implementation supports
16350 this or any other way to make the indexed addressing mode safe to use in
16351 the 32 bit ABI, the default is -mno-indexed-addressing.
16353 @item -mgettrcost=@var{number}
16354 @opindex mgettrcost=@var{number}
16355 Set the cost assumed for the gettr instruction to @var{number}.
16356 The default is 2 if @option{-mpt-fixed} is in effect, 100 otherwise.
16360 Assume pt* instructions won't trap. This will generally generate better
16361 scheduled code, but is unsafe on current hardware. The current architecture
16362 definition says that ptabs and ptrel trap when the target anded with 3 is 3.
16363 This has the unintentional effect of making it unsafe to schedule ptabs /
16364 ptrel before a branch, or hoist it out of a loop. For example,
16365 __do_global_ctors, a part of libgcc that runs constructors at program
16366 startup, calls functions in a list which is delimited by @minus{}1. With the
16367 -mpt-fixed option, the ptabs will be done before testing against @minus{}1.
16368 That means that all the constructors will be run a bit quicker, but when
16369 the loop comes to the end of the list, the program crashes because ptabs
16370 loads @minus{}1 into a target register. Since this option is unsafe for any
16371 hardware implementing the current architecture specification, the default
16372 is -mno-pt-fixed. Unless the user specifies a specific cost with
16373 @option{-mgettrcost}, -mno-pt-fixed also implies @option{-mgettrcost=100};
16374 this deters register allocation using target registers for storing
16377 @item -minvalid-symbols
16378 @opindex minvalid-symbols
16379 Assume symbols might be invalid. Ordinary function symbols generated by
16380 the compiler will always be valid to load with movi/shori/ptabs or
16381 movi/shori/ptrel, but with assembler and/or linker tricks it is possible
16382 to generate symbols that will cause ptabs / ptrel to trap.
16383 This option is only meaningful when @option{-mno-pt-fixed} is in effect.
16384 It will then prevent cross-basic-block cse, hoisting and most scheduling
16385 of symbol loads. The default is @option{-mno-invalid-symbols}.
16388 @node SPARC Options
16389 @subsection SPARC Options
16390 @cindex SPARC options
16392 These @samp{-m} options are supported on the SPARC:
16395 @item -mno-app-regs
16397 @opindex mno-app-regs
16399 Specify @option{-mapp-regs} to generate output using the global registers
16400 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
16403 To be fully SVR4 ABI compliant at the cost of some performance loss,
16404 specify @option{-mno-app-regs}. You should compile libraries and system
16405 software with this option.
16408 @itemx -mhard-float
16410 @opindex mhard-float
16411 Generate output containing floating point instructions. This is the
16415 @itemx -msoft-float
16417 @opindex msoft-float
16418 Generate output containing library calls for floating point.
16419 @strong{Warning:} the requisite libraries are not available for all SPARC
16420 targets. Normally the facilities of the machine's usual C compiler are
16421 used, but this cannot be done directly in cross-compilation. You must make
16422 your own arrangements to provide suitable library functions for
16423 cross-compilation. The embedded targets @samp{sparc-*-aout} and
16424 @samp{sparclite-*-*} do provide software floating point support.
16426 @option{-msoft-float} changes the calling convention in the output file;
16427 therefore, it is only useful if you compile @emph{all} of a program with
16428 this option. In particular, you need to compile @file{libgcc.a}, the
16429 library that comes with GCC, with @option{-msoft-float} in order for
16432 @item -mhard-quad-float
16433 @opindex mhard-quad-float
16434 Generate output containing quad-word (long double) floating point
16437 @item -msoft-quad-float
16438 @opindex msoft-quad-float
16439 Generate output containing library calls for quad-word (long double)
16440 floating point instructions. The functions called are those specified
16441 in the SPARC ABI@. This is the default.
16443 As of this writing, there are no SPARC implementations that have hardware
16444 support for the quad-word floating point instructions. They all invoke
16445 a trap handler for one of these instructions, and then the trap handler
16446 emulates the effect of the instruction. Because of the trap handler overhead,
16447 this is much slower than calling the ABI library routines. Thus the
16448 @option{-msoft-quad-float} option is the default.
16450 @item -mno-unaligned-doubles
16451 @itemx -munaligned-doubles
16452 @opindex mno-unaligned-doubles
16453 @opindex munaligned-doubles
16454 Assume that doubles have 8 byte alignment. This is the default.
16456 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
16457 alignment only if they are contained in another type, or if they have an
16458 absolute address. Otherwise, it assumes they have 4 byte alignment.
16459 Specifying this option avoids some rare compatibility problems with code
16460 generated by other compilers. It is not the default because it results
16461 in a performance loss, especially for floating point code.
16463 @item -mno-faster-structs
16464 @itemx -mfaster-structs
16465 @opindex mno-faster-structs
16466 @opindex mfaster-structs
16467 With @option{-mfaster-structs}, the compiler assumes that structures
16468 should have 8 byte alignment. This enables the use of pairs of
16469 @code{ldd} and @code{std} instructions for copies in structure
16470 assignment, in place of twice as many @code{ld} and @code{st} pairs.
16471 However, the use of this changed alignment directly violates the SPARC
16472 ABI@. Thus, it's intended only for use on targets where the developer
16473 acknowledges that their resulting code will not be directly in line with
16474 the rules of the ABI@.
16476 @item -mimpure-text
16477 @opindex mimpure-text
16478 @option{-mimpure-text}, used in addition to @option{-shared}, tells
16479 the compiler to not pass @option{-z text} to the linker when linking a
16480 shared object. Using this option, you can link position-dependent
16481 code into a shared object.
16483 @option{-mimpure-text} suppresses the ``relocations remain against
16484 allocatable but non-writable sections'' linker error message.
16485 However, the necessary relocations will trigger copy-on-write, and the
16486 shared object is not actually shared across processes. Instead of
16487 using @option{-mimpure-text}, you should compile all source code with
16488 @option{-fpic} or @option{-fPIC}.
16490 This option is only available on SunOS and Solaris.
16492 @item -mcpu=@var{cpu_type}
16494 Set the instruction set, register set, and instruction scheduling parameters
16495 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
16496 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
16497 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
16498 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc},
16499 @samp{ultrasparc3}, @samp{niagara} and @samp{niagara2}.
16501 Default instruction scheduling parameters are used for values that select
16502 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
16503 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
16505 Here is a list of each supported architecture and their supported
16510 v8: supersparc, hypersparc
16511 sparclite: f930, f934, sparclite86x
16513 v9: ultrasparc, ultrasparc3, niagara, niagara2
16516 By default (unless configured otherwise), GCC generates code for the V7
16517 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
16518 additionally optimizes it for the Cypress CY7C602 chip, as used in the
16519 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
16520 SPARCStation 1, 2, IPX etc.
16522 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
16523 architecture. The only difference from V7 code is that the compiler emits
16524 the integer multiply and integer divide instructions which exist in SPARC-V8
16525 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
16526 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
16529 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
16530 the SPARC architecture. This adds the integer multiply, integer divide step
16531 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
16532 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
16533 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
16534 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
16535 MB86934 chip, which is the more recent SPARClite with FPU@.
16537 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
16538 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
16539 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
16540 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
16541 optimizes it for the TEMIC SPARClet chip.
16543 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
16544 architecture. This adds 64-bit integer and floating-point move instructions,
16545 3 additional floating-point condition code registers and conditional move
16546 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
16547 optimizes it for the Sun UltraSPARC I/II/IIi chips. With
16548 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
16549 Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips. With
16550 @option{-mcpu=niagara}, the compiler additionally optimizes it for
16551 Sun UltraSPARC T1 chips. With @option{-mcpu=niagara2}, the compiler
16552 additionally optimizes it for Sun UltraSPARC T2 chips.
16554 @item -mtune=@var{cpu_type}
16556 Set the instruction scheduling parameters for machine type
16557 @var{cpu_type}, but do not set the instruction set or register set that the
16558 option @option{-mcpu=@var{cpu_type}} would.
16560 The same values for @option{-mcpu=@var{cpu_type}} can be used for
16561 @option{-mtune=@var{cpu_type}}, but the only useful values are those
16562 that select a particular cpu implementation. Those are @samp{cypress},
16563 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
16564 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc},
16565 @samp{ultrasparc3}, @samp{niagara}, and @samp{niagara2}.
16570 @opindex mno-v8plus
16571 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
16572 difference from the V8 ABI is that the global and out registers are
16573 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
16574 mode for all SPARC-V9 processors.
16580 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
16581 Visual Instruction Set extensions. The default is @option{-mno-vis}.
16584 These @samp{-m} options are supported in addition to the above
16585 on SPARC-V9 processors in 64-bit environments:
16588 @item -mlittle-endian
16589 @opindex mlittle-endian
16590 Generate code for a processor running in little-endian mode. It is only
16591 available for a few configurations and most notably not on Solaris and Linux.
16597 Generate code for a 32-bit or 64-bit environment.
16598 The 32-bit environment sets int, long and pointer to 32 bits.
16599 The 64-bit environment sets int to 32 bits and long and pointer
16602 @item -mcmodel=medlow
16603 @opindex mcmodel=medlow
16604 Generate code for the Medium/Low code model: 64-bit addresses, programs
16605 must be linked in the low 32 bits of memory. Programs can be statically
16606 or dynamically linked.
16608 @item -mcmodel=medmid
16609 @opindex mcmodel=medmid
16610 Generate code for the Medium/Middle code model: 64-bit addresses, programs
16611 must be linked in the low 44 bits of memory, the text and data segments must
16612 be less than 2GB in size and the data segment must be located within 2GB of
16615 @item -mcmodel=medany
16616 @opindex mcmodel=medany
16617 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
16618 may be linked anywhere in memory, the text and data segments must be less
16619 than 2GB in size and the data segment must be located within 2GB of the
16622 @item -mcmodel=embmedany
16623 @opindex mcmodel=embmedany
16624 Generate code for the Medium/Anywhere code model for embedded systems:
16625 64-bit addresses, the text and data segments must be less than 2GB in
16626 size, both starting anywhere in memory (determined at link time). The
16627 global register %g4 points to the base of the data segment. Programs
16628 are statically linked and PIC is not supported.
16631 @itemx -mno-stack-bias
16632 @opindex mstack-bias
16633 @opindex mno-stack-bias
16634 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
16635 frame pointer if present, are offset by @minus{}2047 which must be added back
16636 when making stack frame references. This is the default in 64-bit mode.
16637 Otherwise, assume no such offset is present.
16640 These switches are supported in addition to the above on Solaris:
16645 Add support for multithreading using the Solaris threads library. This
16646 option sets flags for both the preprocessor and linker. This option does
16647 not affect the thread safety of object code produced by the compiler or
16648 that of libraries supplied with it.
16652 Add support for multithreading using the POSIX threads library. This
16653 option sets flags for both the preprocessor and linker. This option does
16654 not affect the thread safety of object code produced by the compiler or
16655 that of libraries supplied with it.
16659 This is a synonym for @option{-pthreads}.
16663 @subsection SPU Options
16664 @cindex SPU options
16666 These @samp{-m} options are supported on the SPU:
16670 @itemx -merror-reloc
16671 @opindex mwarn-reloc
16672 @opindex merror-reloc
16674 The loader for SPU does not handle dynamic relocations. By default, GCC
16675 will give an error when it generates code that requires a dynamic
16676 relocation. @option{-mno-error-reloc} disables the error,
16677 @option{-mwarn-reloc} will generate a warning instead.
16680 @itemx -munsafe-dma
16682 @opindex munsafe-dma
16684 Instructions which initiate or test completion of DMA must not be
16685 reordered with respect to loads and stores of the memory which is being
16686 accessed. Users typically address this problem using the volatile
16687 keyword, but that can lead to inefficient code in places where the
16688 memory is known to not change. Rather than mark the memory as volatile
16689 we treat the DMA instructions as potentially effecting all memory. With
16690 @option{-munsafe-dma} users must use the volatile keyword to protect
16693 @item -mbranch-hints
16694 @opindex mbranch-hints
16696 By default, GCC will generate a branch hint instruction to avoid
16697 pipeline stalls for always taken or probably taken branches. A hint
16698 will not be generated closer than 8 instructions away from its branch.
16699 There is little reason to disable them, except for debugging purposes,
16700 or to make an object a little bit smaller.
16704 @opindex msmall-mem
16705 @opindex mlarge-mem
16707 By default, GCC generates code assuming that addresses are never larger
16708 than 18 bits. With @option{-mlarge-mem} code is generated that assumes
16709 a full 32 bit address.
16714 By default, GCC links against startup code that assumes the SPU-style
16715 main function interface (which has an unconventional parameter list).
16716 With @option{-mstdmain}, GCC will link your program against startup
16717 code that assumes a C99-style interface to @code{main}, including a
16718 local copy of @code{argv} strings.
16720 @item -mfixed-range=@var{register-range}
16721 @opindex mfixed-range
16722 Generate code treating the given register range as fixed registers.
16723 A fixed register is one that the register allocator can not use. This is
16724 useful when compiling kernel code. A register range is specified as
16725 two registers separated by a dash. Multiple register ranges can be
16726 specified separated by a comma.
16732 Compile code assuming that pointers to the PPU address space accessed
16733 via the @code{__ea} named address space qualifier are either 32 or 64
16734 bits wide. The default is 32 bits. As this is an ABI changing option,
16735 all object code in an executable must be compiled with the same setting.
16737 @item -maddress-space-conversion
16738 @itemx -mno-address-space-conversion
16739 @opindex maddress-space-conversion
16740 @opindex mno-address-space-conversion
16741 Allow/disallow treating the @code{__ea} address space as superset
16742 of the generic address space. This enables explicit type casts
16743 between @code{__ea} and generic pointer as well as implicit
16744 conversions of generic pointers to @code{__ea} pointers. The
16745 default is to allow address space pointer conversions.
16747 @item -mcache-size=@var{cache-size}
16748 @opindex mcache-size
16749 This option controls the version of libgcc that the compiler links to an
16750 executable and selects a software-managed cache for accessing variables
16751 in the @code{__ea} address space with a particular cache size. Possible
16752 options for @var{cache-size} are @samp{8}, @samp{16}, @samp{32}, @samp{64}
16753 and @samp{128}. The default cache size is 64KB.
16755 @item -matomic-updates
16756 @itemx -mno-atomic-updates
16757 @opindex matomic-updates
16758 @opindex mno-atomic-updates
16759 This option controls the version of libgcc that the compiler links to an
16760 executable and selects whether atomic updates to the software-managed
16761 cache of PPU-side variables are used. If you use atomic updates, changes
16762 to a PPU variable from SPU code using the @code{__ea} named address space
16763 qualifier will not interfere with changes to other PPU variables residing
16764 in the same cache line from PPU code. If you do not use atomic updates,
16765 such interference may occur; however, writing back cache lines will be
16766 more efficient. The default behavior is to use atomic updates.
16769 @itemx -mdual-nops=@var{n}
16770 @opindex mdual-nops
16771 By default, GCC will insert nops to increase dual issue when it expects
16772 it to increase performance. @var{n} can be a value from 0 to 10. A
16773 smaller @var{n} will insert fewer nops. 10 is the default, 0 is the
16774 same as @option{-mno-dual-nops}. Disabled with @option{-Os}.
16776 @item -mhint-max-nops=@var{n}
16777 @opindex mhint-max-nops
16778 Maximum number of nops to insert for a branch hint. A branch hint must
16779 be at least 8 instructions away from the branch it is effecting. GCC
16780 will insert up to @var{n} nops to enforce this, otherwise it will not
16781 generate the branch hint.
16783 @item -mhint-max-distance=@var{n}
16784 @opindex mhint-max-distance
16785 The encoding of the branch hint instruction limits the hint to be within
16786 256 instructions of the branch it is effecting. By default, GCC makes
16787 sure it is within 125.
16790 @opindex msafe-hints
16791 Work around a hardware bug which causes the SPU to stall indefinitely.
16792 By default, GCC will insert the @code{hbrp} instruction to make sure
16793 this stall won't happen.
16797 @node System V Options
16798 @subsection Options for System V
16800 These additional options are available on System V Release 4 for
16801 compatibility with other compilers on those systems:
16806 Create a shared object.
16807 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
16811 Identify the versions of each tool used by the compiler, in a
16812 @code{.ident} assembler directive in the output.
16816 Refrain from adding @code{.ident} directives to the output file (this is
16819 @item -YP,@var{dirs}
16821 Search the directories @var{dirs}, and no others, for libraries
16822 specified with @option{-l}.
16824 @item -Ym,@var{dir}
16826 Look in the directory @var{dir} to find the M4 preprocessor.
16827 The assembler uses this option.
16828 @c This is supposed to go with a -Yd for predefined M4 macro files, but
16829 @c the generic assembler that comes with Solaris takes just -Ym.
16833 @subsection V850 Options
16834 @cindex V850 Options
16836 These @samp{-m} options are defined for V850 implementations:
16840 @itemx -mno-long-calls
16841 @opindex mlong-calls
16842 @opindex mno-long-calls
16843 Treat all calls as being far away (near). If calls are assumed to be
16844 far away, the compiler will always load the functions address up into a
16845 register, and call indirect through the pointer.
16851 Do not optimize (do optimize) basic blocks that use the same index
16852 pointer 4 or more times to copy pointer into the @code{ep} register, and
16853 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
16854 option is on by default if you optimize.
16856 @item -mno-prolog-function
16857 @itemx -mprolog-function
16858 @opindex mno-prolog-function
16859 @opindex mprolog-function
16860 Do not use (do use) external functions to save and restore registers
16861 at the prologue and epilogue of a function. The external functions
16862 are slower, but use less code space if more than one function saves
16863 the same number of registers. The @option{-mprolog-function} option
16864 is on by default if you optimize.
16868 Try to make the code as small as possible. At present, this just turns
16869 on the @option{-mep} and @option{-mprolog-function} options.
16871 @item -mtda=@var{n}
16873 Put static or global variables whose size is @var{n} bytes or less into
16874 the tiny data area that register @code{ep} points to. The tiny data
16875 area can hold up to 256 bytes in total (128 bytes for byte references).
16877 @item -msda=@var{n}
16879 Put static or global variables whose size is @var{n} bytes or less into
16880 the small data area that register @code{gp} points to. The small data
16881 area can hold up to 64 kilobytes.
16883 @item -mzda=@var{n}
16885 Put static or global variables whose size is @var{n} bytes or less into
16886 the first 32 kilobytes of memory.
16890 Specify that the target processor is the V850.
16893 @opindex mbig-switch
16894 Generate code suitable for big switch tables. Use this option only if
16895 the assembler/linker complain about out of range branches within a switch
16900 This option will cause r2 and r5 to be used in the code generated by
16901 the compiler. This setting is the default.
16903 @item -mno-app-regs
16904 @opindex mno-app-regs
16905 This option will cause r2 and r5 to be treated as fixed registers.
16909 Specify that the target processor is the V850E1. The preprocessor
16910 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
16911 this option is used.
16915 Specify that the target processor is the V850E@. The preprocessor
16916 constant @samp{__v850e__} will be defined if this option is used.
16918 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
16919 are defined then a default target processor will be chosen and the
16920 relevant @samp{__v850*__} preprocessor constant will be defined.
16922 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
16923 defined, regardless of which processor variant is the target.
16925 @item -mdisable-callt
16926 @opindex mdisable-callt
16927 This option will suppress generation of the CALLT instruction for the
16928 v850e and v850e1 flavors of the v850 architecture. The default is
16929 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
16934 @subsection VAX Options
16935 @cindex VAX options
16937 These @samp{-m} options are defined for the VAX:
16942 Do not output certain jump instructions (@code{aobleq} and so on)
16943 that the Unix assembler for the VAX cannot handle across long
16948 Do output those jump instructions, on the assumption that you
16949 will assemble with the GNU assembler.
16953 Output code for g-format floating point numbers instead of d-format.
16956 @node VxWorks Options
16957 @subsection VxWorks Options
16958 @cindex VxWorks Options
16960 The options in this section are defined for all VxWorks targets.
16961 Options specific to the target hardware are listed with the other
16962 options for that target.
16967 GCC can generate code for both VxWorks kernels and real time processes
16968 (RTPs). This option switches from the former to the latter. It also
16969 defines the preprocessor macro @code{__RTP__}.
16972 @opindex non-static
16973 Link an RTP executable against shared libraries rather than static
16974 libraries. The options @option{-static} and @option{-shared} can
16975 also be used for RTPs (@pxref{Link Options}); @option{-static}
16982 These options are passed down to the linker. They are defined for
16983 compatibility with Diab.
16986 @opindex Xbind-lazy
16987 Enable lazy binding of function calls. This option is equivalent to
16988 @option{-Wl,-z,now} and is defined for compatibility with Diab.
16992 Disable lazy binding of function calls. This option is the default and
16993 is defined for compatibility with Diab.
16996 @node x86-64 Options
16997 @subsection x86-64 Options
16998 @cindex x86-64 options
17000 These are listed under @xref{i386 and x86-64 Options}.
17002 @node i386 and x86-64 Windows Options
17003 @subsection i386 and x86-64 Windows Options
17004 @cindex i386 and x86-64 Windows Options
17006 These additional options are available for Windows targets:
17011 This option is available for Cygwin and MinGW targets. It
17012 specifies that a console application is to be generated, by
17013 instructing the linker to set the PE header subsystem type
17014 required for console applications.
17015 This is the default behavior for Cygwin and MinGW targets.
17019 This option is available for Cygwin targets. It specifies that
17020 the Cygwin internal interface is to be used for predefined
17021 preprocessor macros, C runtime libraries and related linker
17022 paths and options. For Cygwin targets this is the default behavior.
17023 This option is deprecated and will be removed in a future release.
17026 @opindex mno-cygwin
17027 This option is available for Cygwin targets. It specifies that
17028 the MinGW internal interface is to be used instead of Cygwin's, by
17029 setting MinGW-related predefined macros and linker paths and default
17031 This option is deprecated and will be removed in a future release.
17035 This option is available for Cygwin and MinGW targets. It
17036 specifies that a DLL - a dynamic link library - is to be
17037 generated, enabling the selection of the required runtime
17038 startup object and entry point.
17040 @item -mnop-fun-dllimport
17041 @opindex mnop-fun-dllimport
17042 This option is available for Cygwin and MinGW targets. It
17043 specifies that the dllimport attribute should be ignored.
17047 This option is available for MinGW targets. It specifies
17048 that MinGW-specific thread support is to be used.
17052 This option is available for mingw-w64 targets. It specifies
17053 that the UNICODE macro is getting pre-defined and that the
17054 unicode capable runtime startup code is chosen.
17058 This option is available for Cygwin and MinGW targets. It
17059 specifies that the typical Windows pre-defined macros are to
17060 be set in the pre-processor, but does not influence the choice
17061 of runtime library/startup code.
17065 This option is available for Cygwin and MinGW targets. It
17066 specifies that a GUI application is to be generated by
17067 instructing the linker to set the PE header subsystem type
17070 @item -fno-set-stack-executable
17071 @opindex fno-set-stack-executable
17072 This option is available for MinGW targets. It specifies that
17073 the executable flag for stack used by nested functions isn't
17074 set. This is necessary for binaries running in kernel mode of
17075 Windows, as there the user32 API, which is used to set executable
17076 privileges, isn't available.
17078 @item -mpe-aligned-commons
17079 @opindex mpe-aligned-commons
17080 This option is available for Cygwin and MinGW targets. It
17081 specifies that the GNU extension to the PE file format that
17082 permits the correct alignment of COMMON variables should be
17083 used when generating code. It will be enabled by default if
17084 GCC detects that the target assembler found during configuration
17085 supports the feature.
17088 See also under @ref{i386 and x86-64 Options} for standard options.
17090 @node Xstormy16 Options
17091 @subsection Xstormy16 Options
17092 @cindex Xstormy16 Options
17094 These options are defined for Xstormy16:
17099 Choose startup files and linker script suitable for the simulator.
17102 @node Xtensa Options
17103 @subsection Xtensa Options
17104 @cindex Xtensa Options
17106 These options are supported for Xtensa targets:
17110 @itemx -mno-const16
17112 @opindex mno-const16
17113 Enable or disable use of @code{CONST16} instructions for loading
17114 constant values. The @code{CONST16} instruction is currently not a
17115 standard option from Tensilica. When enabled, @code{CONST16}
17116 instructions are always used in place of the standard @code{L32R}
17117 instructions. The use of @code{CONST16} is enabled by default only if
17118 the @code{L32R} instruction is not available.
17121 @itemx -mno-fused-madd
17122 @opindex mfused-madd
17123 @opindex mno-fused-madd
17124 Enable or disable use of fused multiply/add and multiply/subtract
17125 instructions in the floating-point option. This has no effect if the
17126 floating-point option is not also enabled. Disabling fused multiply/add
17127 and multiply/subtract instructions forces the compiler to use separate
17128 instructions for the multiply and add/subtract operations. This may be
17129 desirable in some cases where strict IEEE 754-compliant results are
17130 required: the fused multiply add/subtract instructions do not round the
17131 intermediate result, thereby producing results with @emph{more} bits of
17132 precision than specified by the IEEE standard. Disabling fused multiply
17133 add/subtract instructions also ensures that the program output is not
17134 sensitive to the compiler's ability to combine multiply and add/subtract
17137 @item -mserialize-volatile
17138 @itemx -mno-serialize-volatile
17139 @opindex mserialize-volatile
17140 @opindex mno-serialize-volatile
17141 When this option is enabled, GCC inserts @code{MEMW} instructions before
17142 @code{volatile} memory references to guarantee sequential consistency.
17143 The default is @option{-mserialize-volatile}. Use
17144 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
17146 @item -mforce-no-pic
17147 @opindex mforce-no-pic
17148 For targets, like GNU/Linux, where all user-mode Xtensa code must be
17149 position-independent code (PIC), this option disables PIC for compiling
17152 @item -mtext-section-literals
17153 @itemx -mno-text-section-literals
17154 @opindex mtext-section-literals
17155 @opindex mno-text-section-literals
17156 Control the treatment of literal pools. The default is
17157 @option{-mno-text-section-literals}, which places literals in a separate
17158 section in the output file. This allows the literal pool to be placed
17159 in a data RAM/ROM, and it also allows the linker to combine literal
17160 pools from separate object files to remove redundant literals and
17161 improve code size. With @option{-mtext-section-literals}, the literals
17162 are interspersed in the text section in order to keep them as close as
17163 possible to their references. This may be necessary for large assembly
17166 @item -mtarget-align
17167 @itemx -mno-target-align
17168 @opindex mtarget-align
17169 @opindex mno-target-align
17170 When this option is enabled, GCC instructs the assembler to
17171 automatically align instructions to reduce branch penalties at the
17172 expense of some code density. The assembler attempts to widen density
17173 instructions to align branch targets and the instructions following call
17174 instructions. If there are not enough preceding safe density
17175 instructions to align a target, no widening will be performed. The
17176 default is @option{-mtarget-align}. These options do not affect the
17177 treatment of auto-aligned instructions like @code{LOOP}, which the
17178 assembler will always align, either by widening density instructions or
17179 by inserting no-op instructions.
17182 @itemx -mno-longcalls
17183 @opindex mlongcalls
17184 @opindex mno-longcalls
17185 When this option is enabled, GCC instructs the assembler to translate
17186 direct calls to indirect calls unless it can determine that the target
17187 of a direct call is in the range allowed by the call instruction. This
17188 translation typically occurs for calls to functions in other source
17189 files. Specifically, the assembler translates a direct @code{CALL}
17190 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
17191 The default is @option{-mno-longcalls}. This option should be used in
17192 programs where the call target can potentially be out of range. This
17193 option is implemented in the assembler, not the compiler, so the
17194 assembly code generated by GCC will still show direct call
17195 instructions---look at the disassembled object code to see the actual
17196 instructions. Note that the assembler will use an indirect call for
17197 every cross-file call, not just those that really will be out of range.
17200 @node zSeries Options
17201 @subsection zSeries Options
17202 @cindex zSeries options
17204 These are listed under @xref{S/390 and zSeries Options}.
17206 @node Code Gen Options
17207 @section Options for Code Generation Conventions
17208 @cindex code generation conventions
17209 @cindex options, code generation
17210 @cindex run-time options
17212 These machine-independent options control the interface conventions
17213 used in code generation.
17215 Most of them have both positive and negative forms; the negative form
17216 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
17217 one of the forms is listed---the one which is not the default. You
17218 can figure out the other form by either removing @samp{no-} or adding
17222 @item -fbounds-check
17223 @opindex fbounds-check
17224 For front-ends that support it, generate additional code to check that
17225 indices used to access arrays are within the declared range. This is
17226 currently only supported by the Java and Fortran front-ends, where
17227 this option defaults to true and false respectively.
17231 This option generates traps for signed overflow on addition, subtraction,
17232 multiplication operations.
17236 This option instructs the compiler to assume that signed arithmetic
17237 overflow of addition, subtraction and multiplication wraps around
17238 using twos-complement representation. This flag enables some optimizations
17239 and disables others. This option is enabled by default for the Java
17240 front-end, as required by the Java language specification.
17243 @opindex fexceptions
17244 Enable exception handling. Generates extra code needed to propagate
17245 exceptions. For some targets, this implies GCC will generate frame
17246 unwind information for all functions, which can produce significant data
17247 size overhead, although it does not affect execution. If you do not
17248 specify this option, GCC will enable it by default for languages like
17249 C++ which normally require exception handling, and disable it for
17250 languages like C that do not normally require it. However, you may need
17251 to enable this option when compiling C code that needs to interoperate
17252 properly with exception handlers written in C++. You may also wish to
17253 disable this option if you are compiling older C++ programs that don't
17254 use exception handling.
17256 @item -fnon-call-exceptions
17257 @opindex fnon-call-exceptions
17258 Generate code that allows trapping instructions to throw exceptions.
17259 Note that this requires platform-specific runtime support that does
17260 not exist everywhere. Moreover, it only allows @emph{trapping}
17261 instructions to throw exceptions, i.e.@: memory references or floating
17262 point instructions. It does not allow exceptions to be thrown from
17263 arbitrary signal handlers such as @code{SIGALRM}.
17265 @item -funwind-tables
17266 @opindex funwind-tables
17267 Similar to @option{-fexceptions}, except that it will just generate any needed
17268 static data, but will not affect the generated code in any other way.
17269 You will normally not enable this option; instead, a language processor
17270 that needs this handling would enable it on your behalf.
17272 @item -fasynchronous-unwind-tables
17273 @opindex fasynchronous-unwind-tables
17274 Generate unwind table in dwarf2 format, if supported by target machine. The
17275 table is exact at each instruction boundary, so it can be used for stack
17276 unwinding from asynchronous events (such as debugger or garbage collector).
17278 @item -fpcc-struct-return
17279 @opindex fpcc-struct-return
17280 Return ``short'' @code{struct} and @code{union} values in memory like
17281 longer ones, rather than in registers. This convention is less
17282 efficient, but it has the advantage of allowing intercallability between
17283 GCC-compiled files and files compiled with other compilers, particularly
17284 the Portable C Compiler (pcc).
17286 The precise convention for returning structures in memory depends
17287 on the target configuration macros.
17289 Short structures and unions are those whose size and alignment match
17290 that of some integer type.
17292 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
17293 switch is not binary compatible with code compiled with the
17294 @option{-freg-struct-return} switch.
17295 Use it to conform to a non-default application binary interface.
17297 @item -freg-struct-return
17298 @opindex freg-struct-return
17299 Return @code{struct} and @code{union} values in registers when possible.
17300 This is more efficient for small structures than
17301 @option{-fpcc-struct-return}.
17303 If you specify neither @option{-fpcc-struct-return} nor
17304 @option{-freg-struct-return}, GCC defaults to whichever convention is
17305 standard for the target. If there is no standard convention, GCC
17306 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
17307 the principal compiler. In those cases, we can choose the standard, and
17308 we chose the more efficient register return alternative.
17310 @strong{Warning:} code compiled with the @option{-freg-struct-return}
17311 switch is not binary compatible with code compiled with the
17312 @option{-fpcc-struct-return} switch.
17313 Use it to conform to a non-default application binary interface.
17315 @item -fshort-enums
17316 @opindex fshort-enums
17317 Allocate to an @code{enum} type only as many bytes as it needs for the
17318 declared range of possible values. Specifically, the @code{enum} type
17319 will be equivalent to the smallest integer type which has enough room.
17321 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
17322 code that is not binary compatible with code generated without that switch.
17323 Use it to conform to a non-default application binary interface.
17325 @item -fshort-double
17326 @opindex fshort-double
17327 Use the same size for @code{double} as for @code{float}.
17329 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
17330 code that is not binary compatible with code generated without that switch.
17331 Use it to conform to a non-default application binary interface.
17333 @item -fshort-wchar
17334 @opindex fshort-wchar
17335 Override the underlying type for @samp{wchar_t} to be @samp{short
17336 unsigned int} instead of the default for the target. This option is
17337 useful for building programs to run under WINE@.
17339 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
17340 code that is not binary compatible with code generated without that switch.
17341 Use it to conform to a non-default application binary interface.
17344 @opindex fno-common
17345 In C code, controls the placement of uninitialized global variables.
17346 Unix C compilers have traditionally permitted multiple definitions of
17347 such variables in different compilation units by placing the variables
17349 This is the behavior specified by @option{-fcommon}, and is the default
17350 for GCC on most targets.
17351 On the other hand, this behavior is not required by ISO C, and on some
17352 targets may carry a speed or code size penalty on variable references.
17353 The @option{-fno-common} option specifies that the compiler should place
17354 uninitialized global variables in the data section of the object file,
17355 rather than generating them as common blocks.
17356 This has the effect that if the same variable is declared
17357 (without @code{extern}) in two different compilations,
17358 you will get a multiple-definition error when you link them.
17359 In this case, you must compile with @option{-fcommon} instead.
17360 Compiling with @option{-fno-common} is useful on targets for which
17361 it provides better performance, or if you wish to verify that the
17362 program will work on other systems which always treat uninitialized
17363 variable declarations this way.
17367 Ignore the @samp{#ident} directive.
17369 @item -finhibit-size-directive
17370 @opindex finhibit-size-directive
17371 Don't output a @code{.size} assembler directive, or anything else that
17372 would cause trouble if the function is split in the middle, and the
17373 two halves are placed at locations far apart in memory. This option is
17374 used when compiling @file{crtstuff.c}; you should not need to use it
17377 @item -fverbose-asm
17378 @opindex fverbose-asm
17379 Put extra commentary information in the generated assembly code to
17380 make it more readable. This option is generally only of use to those
17381 who actually need to read the generated assembly code (perhaps while
17382 debugging the compiler itself).
17384 @option{-fno-verbose-asm}, the default, causes the
17385 extra information to be omitted and is useful when comparing two assembler
17388 @item -frecord-gcc-switches
17389 @opindex frecord-gcc-switches
17390 This switch causes the command line that was used to invoke the
17391 compiler to be recorded into the object file that is being created.
17392 This switch is only implemented on some targets and the exact format
17393 of the recording is target and binary file format dependent, but it
17394 usually takes the form of a section containing ASCII text. This
17395 switch is related to the @option{-fverbose-asm} switch, but that
17396 switch only records information in the assembler output file as
17397 comments, so it never reaches the object file.
17401 @cindex global offset table
17403 Generate position-independent code (PIC) suitable for use in a shared
17404 library, if supported for the target machine. Such code accesses all
17405 constant addresses through a global offset table (GOT)@. The dynamic
17406 loader resolves the GOT entries when the program starts (the dynamic
17407 loader is not part of GCC; it is part of the operating system). If
17408 the GOT size for the linked executable exceeds a machine-specific
17409 maximum size, you get an error message from the linker indicating that
17410 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
17411 instead. (These maximums are 8k on the SPARC and 32k
17412 on the m68k and RS/6000. The 386 has no such limit.)
17414 Position-independent code requires special support, and therefore works
17415 only on certain machines. For the 386, GCC supports PIC for System V
17416 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
17417 position-independent.
17419 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
17424 If supported for the target machine, emit position-independent code,
17425 suitable for dynamic linking and avoiding any limit on the size of the
17426 global offset table. This option makes a difference on the m68k,
17427 PowerPC and SPARC@.
17429 Position-independent code requires special support, and therefore works
17430 only on certain machines.
17432 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
17439 These options are similar to @option{-fpic} and @option{-fPIC}, but
17440 generated position independent code can be only linked into executables.
17441 Usually these options are used when @option{-pie} GCC option will be
17442 used during linking.
17444 @option{-fpie} and @option{-fPIE} both define the macros
17445 @code{__pie__} and @code{__PIE__}. The macros have the value 1
17446 for @option{-fpie} and 2 for @option{-fPIE}.
17448 @item -fno-jump-tables
17449 @opindex fno-jump-tables
17450 Do not use jump tables for switch statements even where it would be
17451 more efficient than other code generation strategies. This option is
17452 of use in conjunction with @option{-fpic} or @option{-fPIC} for
17453 building code which forms part of a dynamic linker and cannot
17454 reference the address of a jump table. On some targets, jump tables
17455 do not require a GOT and this option is not needed.
17457 @item -ffixed-@var{reg}
17459 Treat the register named @var{reg} as a fixed register; generated code
17460 should never refer to it (except perhaps as a stack pointer, frame
17461 pointer or in some other fixed role).
17463 @var{reg} must be the name of a register. The register names accepted
17464 are machine-specific and are defined in the @code{REGISTER_NAMES}
17465 macro in the machine description macro file.
17467 This flag does not have a negative form, because it specifies a
17470 @item -fcall-used-@var{reg}
17471 @opindex fcall-used
17472 Treat the register named @var{reg} as an allocable register that is
17473 clobbered by function calls. It may be allocated for temporaries or
17474 variables that do not live across a call. Functions compiled this way
17475 will not save and restore the register @var{reg}.
17477 It is an error to used this flag with the frame pointer or stack pointer.
17478 Use of this flag for other registers that have fixed pervasive roles in
17479 the machine's execution model will produce disastrous results.
17481 This flag does not have a negative form, because it specifies a
17484 @item -fcall-saved-@var{reg}
17485 @opindex fcall-saved
17486 Treat the register named @var{reg} as an allocable register saved by
17487 functions. It may be allocated even for temporaries or variables that
17488 live across a call. Functions compiled this way will save and restore
17489 the register @var{reg} if they use it.
17491 It is an error to used this flag with the frame pointer or stack pointer.
17492 Use of this flag for other registers that have fixed pervasive roles in
17493 the machine's execution model will produce disastrous results.
17495 A different sort of disaster will result from the use of this flag for
17496 a register in which function values may be returned.
17498 This flag does not have a negative form, because it specifies a
17501 @item -fpack-struct[=@var{n}]
17502 @opindex fpack-struct
17503 Without a value specified, pack all structure members together without
17504 holes. When a value is specified (which must be a small power of two), pack
17505 structure members according to this value, representing the maximum
17506 alignment (that is, objects with default alignment requirements larger than
17507 this will be output potentially unaligned at the next fitting location.
17509 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
17510 code that is not binary compatible with code generated without that switch.
17511 Additionally, it makes the code suboptimal.
17512 Use it to conform to a non-default application binary interface.
17514 @item -finstrument-functions
17515 @opindex finstrument-functions
17516 Generate instrumentation calls for entry and exit to functions. Just
17517 after function entry and just before function exit, the following
17518 profiling functions will be called with the address of the current
17519 function and its call site. (On some platforms,
17520 @code{__builtin_return_address} does not work beyond the current
17521 function, so the call site information may not be available to the
17522 profiling functions otherwise.)
17525 void __cyg_profile_func_enter (void *this_fn,
17527 void __cyg_profile_func_exit (void *this_fn,
17531 The first argument is the address of the start of the current function,
17532 which may be looked up exactly in the symbol table.
17534 This instrumentation is also done for functions expanded inline in other
17535 functions. The profiling calls will indicate where, conceptually, the
17536 inline function is entered and exited. This means that addressable
17537 versions of such functions must be available. If all your uses of a
17538 function are expanded inline, this may mean an additional expansion of
17539 code size. If you use @samp{extern inline} in your C code, an
17540 addressable version of such functions must be provided. (This is
17541 normally the case anyways, but if you get lucky and the optimizer always
17542 expands the functions inline, you might have gotten away without
17543 providing static copies.)
17545 A function may be given the attribute @code{no_instrument_function}, in
17546 which case this instrumentation will not be done. This can be used, for
17547 example, for the profiling functions listed above, high-priority
17548 interrupt routines, and any functions from which the profiling functions
17549 cannot safely be called (perhaps signal handlers, if the profiling
17550 routines generate output or allocate memory).
17552 @item -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}
17553 @opindex finstrument-functions-exclude-file-list
17555 Set the list of functions that are excluded from instrumentation (see
17556 the description of @code{-finstrument-functions}). If the file that
17557 contains a function definition matches with one of @var{file}, then
17558 that function is not instrumented. The match is done on substrings:
17559 if the @var{file} parameter is a substring of the file name, it is
17560 considered to be a match.
17563 @code{-finstrument-functions-exclude-file-list=/bits/stl,include/sys}
17564 will exclude any inline function defined in files whose pathnames
17565 contain @code{/bits/stl} or @code{include/sys}.
17567 If, for some reason, you want to include letter @code{','} in one of
17568 @var{sym}, write @code{'\,'}. For example,
17569 @code{-finstrument-functions-exclude-file-list='\,\,tmp'}
17570 (note the single quote surrounding the option).
17572 @item -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{}
17573 @opindex finstrument-functions-exclude-function-list
17575 This is similar to @code{-finstrument-functions-exclude-file-list},
17576 but this option sets the list of function names to be excluded from
17577 instrumentation. The function name to be matched is its user-visible
17578 name, such as @code{vector<int> blah(const vector<int> &)}, not the
17579 internal mangled name (e.g., @code{_Z4blahRSt6vectorIiSaIiEE}). The
17580 match is done on substrings: if the @var{sym} parameter is a substring
17581 of the function name, it is considered to be a match. For C99 and C++
17582 extended identifiers, the function name must be given in UTF-8, not
17583 using universal character names.
17585 @item -fstack-check
17586 @opindex fstack-check
17587 Generate code to verify that you do not go beyond the boundary of the
17588 stack. You should specify this flag if you are running in an
17589 environment with multiple threads, but only rarely need to specify it in
17590 a single-threaded environment since stack overflow is automatically
17591 detected on nearly all systems if there is only one stack.
17593 Note that this switch does not actually cause checking to be done; the
17594 operating system or the language runtime must do that. The switch causes
17595 generation of code to ensure that they see the stack being extended.
17597 You can additionally specify a string parameter: @code{no} means no
17598 checking, @code{generic} means force the use of old-style checking,
17599 @code{specific} means use the best checking method and is equivalent
17600 to bare @option{-fstack-check}.
17602 Old-style checking is a generic mechanism that requires no specific
17603 target support in the compiler but comes with the following drawbacks:
17607 Modified allocation strategy for large objects: they will always be
17608 allocated dynamically if their size exceeds a fixed threshold.
17611 Fixed limit on the size of the static frame of functions: when it is
17612 topped by a particular function, stack checking is not reliable and
17613 a warning is issued by the compiler.
17616 Inefficiency: because of both the modified allocation strategy and the
17617 generic implementation, the performances of the code are hampered.
17620 Note that old-style stack checking is also the fallback method for
17621 @code{specific} if no target support has been added in the compiler.
17623 @item -fstack-limit-register=@var{reg}
17624 @itemx -fstack-limit-symbol=@var{sym}
17625 @itemx -fno-stack-limit
17626 @opindex fstack-limit-register
17627 @opindex fstack-limit-symbol
17628 @opindex fno-stack-limit
17629 Generate code to ensure that the stack does not grow beyond a certain value,
17630 either the value of a register or the address of a symbol. If the stack
17631 would grow beyond the value, a signal is raised. For most targets,
17632 the signal is raised before the stack overruns the boundary, so
17633 it is possible to catch the signal without taking special precautions.
17635 For instance, if the stack starts at absolute address @samp{0x80000000}
17636 and grows downwards, you can use the flags
17637 @option{-fstack-limit-symbol=__stack_limit} and
17638 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
17639 of 128KB@. Note that this may only work with the GNU linker.
17641 @item -fleading-underscore
17642 @opindex fleading-underscore
17643 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
17644 change the way C symbols are represented in the object file. One use
17645 is to help link with legacy assembly code.
17647 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
17648 generate code that is not binary compatible with code generated without that
17649 switch. Use it to conform to a non-default application binary interface.
17650 Not all targets provide complete support for this switch.
17652 @item -ftls-model=@var{model}
17653 @opindex ftls-model
17654 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
17655 The @var{model} argument should be one of @code{global-dynamic},
17656 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
17658 The default without @option{-fpic} is @code{initial-exec}; with
17659 @option{-fpic} the default is @code{global-dynamic}.
17661 @item -fvisibility=@var{default|internal|hidden|protected}
17662 @opindex fvisibility
17663 Set the default ELF image symbol visibility to the specified option---all
17664 symbols will be marked with this unless overridden within the code.
17665 Using this feature can very substantially improve linking and
17666 load times of shared object libraries, produce more optimized
17667 code, provide near-perfect API export and prevent symbol clashes.
17668 It is @strong{strongly} recommended that you use this in any shared objects
17671 Despite the nomenclature, @code{default} always means public ie;
17672 available to be linked against from outside the shared object.
17673 @code{protected} and @code{internal} are pretty useless in real-world
17674 usage so the only other commonly used option will be @code{hidden}.
17675 The default if @option{-fvisibility} isn't specified is
17676 @code{default}, i.e., make every
17677 symbol public---this causes the same behavior as previous versions of
17680 A good explanation of the benefits offered by ensuring ELF
17681 symbols have the correct visibility is given by ``How To Write
17682 Shared Libraries'' by Ulrich Drepper (which can be found at
17683 @w{@uref{http://people.redhat.com/~drepper/}})---however a superior
17684 solution made possible by this option to marking things hidden when
17685 the default is public is to make the default hidden and mark things
17686 public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden}
17687 and @code{__attribute__ ((visibility("default")))} instead of
17688 @code{__declspec(dllexport)} you get almost identical semantics with
17689 identical syntax. This is a great boon to those working with
17690 cross-platform projects.
17692 For those adding visibility support to existing code, you may find
17693 @samp{#pragma GCC visibility} of use. This works by you enclosing
17694 the declarations you wish to set visibility for with (for example)
17695 @samp{#pragma GCC visibility push(hidden)} and
17696 @samp{#pragma GCC visibility pop}.
17697 Bear in mind that symbol visibility should be viewed @strong{as
17698 part of the API interface contract} and thus all new code should
17699 always specify visibility when it is not the default ie; declarations
17700 only for use within the local DSO should @strong{always} be marked explicitly
17701 as hidden as so to avoid PLT indirection overheads---making this
17702 abundantly clear also aids readability and self-documentation of the code.
17703 Note that due to ISO C++ specification requirements, operator new and
17704 operator delete must always be of default visibility.
17706 Be aware that headers from outside your project, in particular system
17707 headers and headers from any other library you use, may not be
17708 expecting to be compiled with visibility other than the default. You
17709 may need to explicitly say @samp{#pragma GCC visibility push(default)}
17710 before including any such headers.
17712 @samp{extern} declarations are not affected by @samp{-fvisibility}, so
17713 a lot of code can be recompiled with @samp{-fvisibility=hidden} with
17714 no modifications. However, this means that calls to @samp{extern}
17715 functions with no explicit visibility will use the PLT, so it is more
17716 effective to use @samp{__attribute ((visibility))} and/or
17717 @samp{#pragma GCC visibility} to tell the compiler which @samp{extern}
17718 declarations should be treated as hidden.
17720 Note that @samp{-fvisibility} does affect C++ vague linkage
17721 entities. This means that, for instance, an exception class that will
17722 be thrown between DSOs must be explicitly marked with default
17723 visibility so that the @samp{type_info} nodes will be unified between
17726 An overview of these techniques, their benefits and how to use them
17727 is at @w{@uref{http://gcc.gnu.org/wiki/Visibility}}.
17729 @item -fstrict-volatile-bitfields
17730 This option should be used if accesses to volatile bitfields (or other
17731 structure fields, although the compiler usually honors those types
17732 anyway) should use a single access in a mode of the same size as the
17733 container's type, aligned to a natural alignment if possible. For
17734 example, targets with memory-mapped peripheral registers might require
17735 all such accesses to be 16 bits wide; with this flag the user could
17736 declare all peripheral bitfields as ``unsigned short'' (assuming short
17737 is 16 bits on these targets) to force GCC to use 16 bit accesses
17738 instead of, perhaps, a more efficient 32 bit access.
17740 If this option is disabled, the compiler will use the most efficient
17741 instruction. In the previous example, that might be a 32-bit load
17742 instruction, even though that will access bytes that do not contain
17743 any portion of the bitfield, or memory-mapped registers unrelated to
17744 the one being updated.
17746 If the target requires strict alignment, and honoring the container
17747 type would require violating this alignment, a warning is issued.
17748 However, the access happens as the user requested, under the
17749 assumption that the user knows something about the target hardware
17750 that GCC is unaware of.
17752 The default value of this option is determined by the application binary
17753 interface for the target processor.
17759 @node Environment Variables
17760 @section Environment Variables Affecting GCC
17761 @cindex environment variables
17763 @c man begin ENVIRONMENT
17764 This section describes several environment variables that affect how GCC
17765 operates. Some of them work by specifying directories or prefixes to use
17766 when searching for various kinds of files. Some are used to specify other
17767 aspects of the compilation environment.
17769 Note that you can also specify places to search using options such as
17770 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
17771 take precedence over places specified using environment variables, which
17772 in turn take precedence over those specified by the configuration of GCC@.
17773 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
17774 GNU Compiler Collection (GCC) Internals}.
17779 @c @itemx LC_COLLATE
17781 @c @itemx LC_MONETARY
17782 @c @itemx LC_NUMERIC
17787 @c @findex LC_COLLATE
17788 @findex LC_MESSAGES
17789 @c @findex LC_MONETARY
17790 @c @findex LC_NUMERIC
17794 These environment variables control the way that GCC uses
17795 localization information that allow GCC to work with different
17796 national conventions. GCC inspects the locale categories
17797 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
17798 so. These locale categories can be set to any value supported by your
17799 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
17800 Kingdom encoded in UTF-8.
17802 The @env{LC_CTYPE} environment variable specifies character
17803 classification. GCC uses it to determine the character boundaries in
17804 a string; this is needed for some multibyte encodings that contain quote
17805 and escape characters that would otherwise be interpreted as a string
17808 The @env{LC_MESSAGES} environment variable specifies the language to
17809 use in diagnostic messages.
17811 If the @env{LC_ALL} environment variable is set, it overrides the value
17812 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
17813 and @env{LC_MESSAGES} default to the value of the @env{LANG}
17814 environment variable. If none of these variables are set, GCC
17815 defaults to traditional C English behavior.
17819 If @env{TMPDIR} is set, it specifies the directory to use for temporary
17820 files. GCC uses temporary files to hold the output of one stage of
17821 compilation which is to be used as input to the next stage: for example,
17822 the output of the preprocessor, which is the input to the compiler
17825 @item GCC_EXEC_PREFIX
17826 @findex GCC_EXEC_PREFIX
17827 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
17828 names of the subprograms executed by the compiler. No slash is added
17829 when this prefix is combined with the name of a subprogram, but you can
17830 specify a prefix that ends with a slash if you wish.
17832 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
17833 an appropriate prefix to use based on the pathname it was invoked with.
17835 If GCC cannot find the subprogram using the specified prefix, it
17836 tries looking in the usual places for the subprogram.
17838 The default value of @env{GCC_EXEC_PREFIX} is
17839 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the prefix to
17840 the installed compiler. In many cases @var{prefix} is the value
17841 of @code{prefix} when you ran the @file{configure} script.
17843 Other prefixes specified with @option{-B} take precedence over this prefix.
17845 This prefix is also used for finding files such as @file{crt0.o} that are
17848 In addition, the prefix is used in an unusual way in finding the
17849 directories to search for header files. For each of the standard
17850 directories whose name normally begins with @samp{/usr/local/lib/gcc}
17851 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
17852 replacing that beginning with the specified prefix to produce an
17853 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
17854 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
17855 These alternate directories are searched first; the standard directories
17856 come next. If a standard directory begins with the configured
17857 @var{prefix} then the value of @var{prefix} is replaced by
17858 @env{GCC_EXEC_PREFIX} when looking for header files.
17860 @item COMPILER_PATH
17861 @findex COMPILER_PATH
17862 The value of @env{COMPILER_PATH} is a colon-separated list of
17863 directories, much like @env{PATH}. GCC tries the directories thus
17864 specified when searching for subprograms, if it can't find the
17865 subprograms using @env{GCC_EXEC_PREFIX}.
17868 @findex LIBRARY_PATH
17869 The value of @env{LIBRARY_PATH} is a colon-separated list of
17870 directories, much like @env{PATH}. When configured as a native compiler,
17871 GCC tries the directories thus specified when searching for special
17872 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
17873 using GCC also uses these directories when searching for ordinary
17874 libraries for the @option{-l} option (but directories specified with
17875 @option{-L} come first).
17879 @cindex locale definition
17880 This variable is used to pass locale information to the compiler. One way in
17881 which this information is used is to determine the character set to be used
17882 when character literals, string literals and comments are parsed in C and C++.
17883 When the compiler is configured to allow multibyte characters,
17884 the following values for @env{LANG} are recognized:
17888 Recognize JIS characters.
17890 Recognize SJIS characters.
17892 Recognize EUCJP characters.
17895 If @env{LANG} is not defined, or if it has some other value, then the
17896 compiler will use mblen and mbtowc as defined by the default locale to
17897 recognize and translate multibyte characters.
17901 Some additional environments variables affect the behavior of the
17904 @include cppenv.texi
17908 @node Precompiled Headers
17909 @section Using Precompiled Headers
17910 @cindex precompiled headers
17911 @cindex speed of compilation
17913 Often large projects have many header files that are included in every
17914 source file. The time the compiler takes to process these header files
17915 over and over again can account for nearly all of the time required to
17916 build the project. To make builds faster, GCC allows users to
17917 `precompile' a header file; then, if builds can use the precompiled
17918 header file they will be much faster.
17920 To create a precompiled header file, simply compile it as you would any
17921 other file, if necessary using the @option{-x} option to make the driver
17922 treat it as a C or C++ header file. You will probably want to use a
17923 tool like @command{make} to keep the precompiled header up-to-date when
17924 the headers it contains change.
17926 A precompiled header file will be searched for when @code{#include} is
17927 seen in the compilation. As it searches for the included file
17928 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
17929 compiler looks for a precompiled header in each directory just before it
17930 looks for the include file in that directory. The name searched for is
17931 the name specified in the @code{#include} with @samp{.gch} appended. If
17932 the precompiled header file can't be used, it is ignored.
17934 For instance, if you have @code{#include "all.h"}, and you have
17935 @file{all.h.gch} in the same directory as @file{all.h}, then the
17936 precompiled header file will be used if possible, and the original
17937 header will be used otherwise.
17939 Alternatively, you might decide to put the precompiled header file in a
17940 directory and use @option{-I} to ensure that directory is searched
17941 before (or instead of) the directory containing the original header.
17942 Then, if you want to check that the precompiled header file is always
17943 used, you can put a file of the same name as the original header in this
17944 directory containing an @code{#error} command.
17946 This also works with @option{-include}. So yet another way to use
17947 precompiled headers, good for projects not designed with precompiled
17948 header files in mind, is to simply take most of the header files used by
17949 a project, include them from another header file, precompile that header
17950 file, and @option{-include} the precompiled header. If the header files
17951 have guards against multiple inclusion, they will be skipped because
17952 they've already been included (in the precompiled header).
17954 If you need to precompile the same header file for different
17955 languages, targets, or compiler options, you can instead make a
17956 @emph{directory} named like @file{all.h.gch}, and put each precompiled
17957 header in the directory, perhaps using @option{-o}. It doesn't matter
17958 what you call the files in the directory, every precompiled header in
17959 the directory will be considered. The first precompiled header
17960 encountered in the directory that is valid for this compilation will
17961 be used; they're searched in no particular order.
17963 There are many other possibilities, limited only by your imagination,
17964 good sense, and the constraints of your build system.
17966 A precompiled header file can be used only when these conditions apply:
17970 Only one precompiled header can be used in a particular compilation.
17973 A precompiled header can't be used once the first C token is seen. You
17974 can have preprocessor directives before a precompiled header; you can
17975 even include a precompiled header from inside another header, so long as
17976 there are no C tokens before the @code{#include}.
17979 The precompiled header file must be produced for the same language as
17980 the current compilation. You can't use a C precompiled header for a C++
17984 The precompiled header file must have been produced by the same compiler
17985 binary as the current compilation is using.
17988 Any macros defined before the precompiled header is included must
17989 either be defined in the same way as when the precompiled header was
17990 generated, or must not affect the precompiled header, which usually
17991 means that they don't appear in the precompiled header at all.
17993 The @option{-D} option is one way to define a macro before a
17994 precompiled header is included; using a @code{#define} can also do it.
17995 There are also some options that define macros implicitly, like
17996 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
17999 @item If debugging information is output when using the precompiled
18000 header, using @option{-g} or similar, the same kind of debugging information
18001 must have been output when building the precompiled header. However,
18002 a precompiled header built using @option{-g} can be used in a compilation
18003 when no debugging information is being output.
18005 @item The same @option{-m} options must generally be used when building
18006 and using the precompiled header. @xref{Submodel Options},
18007 for any cases where this rule is relaxed.
18009 @item Each of the following options must be the same when building and using
18010 the precompiled header:
18012 @gccoptlist{-fexceptions}
18015 Some other command-line options starting with @option{-f},
18016 @option{-p}, or @option{-O} must be defined in the same way as when
18017 the precompiled header was generated. At present, it's not clear
18018 which options are safe to change and which are not; the safest choice
18019 is to use exactly the same options when generating and using the
18020 precompiled header. The following are known to be safe:
18022 @gccoptlist{-fmessage-length= -fpreprocessed -fsched-interblock @gol
18023 -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous @gol
18024 -fsched-verbose=<number> -fschedule-insns -fvisibility= @gol
18029 For all of these except the last, the compiler will automatically
18030 ignore the precompiled header if the conditions aren't met. If you
18031 find an option combination that doesn't work and doesn't cause the
18032 precompiled header to be ignored, please consider filing a bug report,
18035 If you do use differing options when generating and using the
18036 precompiled header, the actual behavior will be a mixture of the
18037 behavior for the options. For instance, if you use @option{-g} to
18038 generate the precompiled header but not when using it, you may or may
18039 not get debugging information for routines in the precompiled header.